R Programming Text Book

R Programming
– An Approach to Data Analytics
Libros de Estadística-Ciencia de Datos|Statistics-Data Science Books (PDF)

R Programming
– An Approach to Data Analytics
Dr. G. Sudhamathy
Assistant Professor, Department of Computer Science,

Avinashilingam University, Coimbatore – 641043.
Dr. C. Jothi Venkateswaran
Principal, Government Arts and Science College,

Perumbakkam, Chennai – 600100.
Chennai New Delhi Tirunelveli

right
r Copy
Honou & cy
e Pira
Exclud
ed by
rotect
b o o k i s p duction of
This epro
ght. R cluding
c o p y r i in any form in e done
p a r t not b
any shall
opying on from
photoc ith authorizati
w
except sher.
li
the pub
ISBN 978-81-8094-408-6 MJP Publishers

All rights reserved No. 44, Nallathambi Street,
Printed and bound in India Triplicane, Chennai 600 005
MJP 371 © Publishers, 2018
Publisher: C. Janarthanan
Project Editor: C. Ambica

FOREWORD
If you are looking for a complete step-by-step instructions for learning R Programming
for Statistical Data Analysis, Graphical Visualization and Data Mining, authors Dr.
Sudhamathy & Dr. Jothi venkateswaran’s “R Programming - An Approach to Data
Analytics” is a hands-on book packed with examples and references that would help
you get started coding in R for variety of data science problems.
As the authors explain in their book, understanding the techniques and

algorithms of data analytics for large dataset is critical for effective data classification.
This helps as developer not just learn R Programming but also to apply right
algorithms and statistical model.
Hopefully you can take the instructions provided in this book to get started in R
programming for your next data analysis project, do some exciting data visualization
and data mining on your own.
Mr. Sajeev Madhavan

Director of Architecture, ORACLE,
San Francisco Bay Area, USA

FOREWORD
It’s my immense happiness in penning this foreword for a book that is quite impressive
for any techie who is interested in R-programming. It’s also equally joyous to have a
book written by experts, Dr. G. Sudhamathy and Dr. C. Jothi Venkateswaran. When
a book can teach you and guide you as you work hands on the tool, you are in the
right direction in your learning path.
This book on R-Programming starts with simple concepts like programming

logic statements, data types, moves on to cover some advanced topics like Statistical
Analysis using R, Data mining using R, in addition to the Graphics programming.
The concepts are well supported by the real-time works carried out in R, with
sufficient figurative illustrations wherever it is strongly essential for explanations.
The powerful presentations on the Analytics chapters have an in-depth knowledge
of the tool, with a vital emphasis on the Analytics which have been supported with
the five most important case-studies given towards the end of the book.
One can be definitively sure that book will be of great help and guidance for the
learner to carry out their works on Analytics using R, either in the research, practice
or just to learn the tool.
My heartfelt appreciations to the authors who have done a wonderful job in

bringing out this book which is very much needed at this point of the hour where
the entire world is diving into data, Big data and analyzing those data for newer
knowledge and perceptions to drive everyday business.
Best wishes for a bestselling of this book in the Academia, Research and Practice.
Dr. S. Justus
Associate Professor & Chair - Software Engineering Research Group
VIT University, Chennai
PREFACE
Huge volumes of data are being generated by many sources like commercial
enterprises, scientific domains and general public daily. According to a recent
research, data production will be 44 times greater in 2020 than it was in 2010.
Data being a vital resource for business organizations and other domains like
education, health, manufacturing etc., its management and analysis is becoming
increasingly important. This data, due to its volume, variety and velocity, often
referred to as Big Data, also includes highly unstructured data in the form of textual
documents, web pages, graphical information and social media comments. Since
Big Data is characterised by massive sample sizes, high dimensionality and intrinsic
heterogeneity, traditional approaches to data management, visualisation and
analytics are no longer satisfactorily applicable. There is therefore an urgent need
for newer tools, better frameworks and workable methodologies for such data to
be appropriately categorised, logically segmented, efficiently analysed and securely
managed. This requirement has resulted in an emerging new discipline of Data
Science that is now gaining much attention with researchers and practitioners in
the field of Data Analytics.
R is a programming language and a free open source software environment for

data analytics. It is growing exponentially by most measures, count over a million
users, and it has over 10,865 standard and add-on packages contributed by the
community, with that number increasing by about 25% each year. R is a powerful
tool for approaching statistical, graphical, and data mining problems. It is used
by many organizations and individuals daily to perform serious data analytics. R
does not require the users to have basic programming knowledge as it is made up
of many inbuilt packages and function which can achieve very complex processing
easily in fraction of seconds. This book is full of easy simple steps to achieve greater
results with complex data. It also details on how to model a specific problem and
come out with predictions for the future. The main motivation of this book is to
break the complexities remaining in the minds of students and researchers about
x Preface
R programming language and make it easy to approach by any one. The chapters
are designed in such a fashion that it targets the beginners with the first 4 chapters
and targets the advanced concept learners in the next 3 chapters. The book also
helps the reader with the list of all packages and functions used in this book along
with the page numbers to know the usage of those. Every concept discussed in the
various sections in this book has proper example dealt with a set of code and its
results (as text or as graphs).
The book is organized into 7 chapters and the concept discussed in each chapter
is as detailed below.
Chapter 1 introduces the programming language R, briefs on how to install R

Studio, how to use the editor and write simple code using R. This chapter also details
on how to get help in R from its manuals and how to perform simple mathematical
operations using R. The chapter then progresses with the introduction of the
concepts of packages, environments and functions. Finally this chapter details on
the programming concepts of flow control and loops.
Chapter 2 discusses on the basic data types in R, the primitive data types such
as vectors, matrices and arrays, lists and factors. It also deals with the complex data
types such as data frames, strings, dates and times. The chapter not only discusses
on the data creation, but also basic operations on the data of different data types.
Chapter 3 deals with data preparation in which it details on how and where to
fetch the datasets from, how to import and export data from various sources which
are of different types like CSV files, XML files, etc. It also discusses on the ways of
accessing various databases. The data cleaning and transformation techniques such
as data reshaping, grouping functions are also outlined in this chapter.
Chapter 4 is about using the graphical features in R for exploratory data analysis.
It gives examples of pie charts, scatter plots, line plots, histograms, box plots and
bar plots using the various graphical packages such as base, lattice and ggplot2.
Chapter 5 deals with statistical analysis concepts using R such as the basic
statistical measures like mean, median, mode, standard deviation, variance and
ranges. It discusses on the distribution of data as normal distribution and binomial

xi Preface
distribution and how it can be viewed and analyzed using R. Then, the chapter
explores on the complex statistical techniques such as correlation analysis, regression
analysis, ANOVA and hypothesis testing which can be implemented using R.
Chapter 6 is about exploring the data mining techniques available in R. It

explores the K-Means, K-Medoids, Hierarchical and Density Based Clustering
techniques using proper examples and case studies. The decision tree classification
techniques are also discussed with suitable examples. Outlier detection is also
explored using various techniques such as univariate, multivariate, LOF and
clustering. Dimensionality reduction is done using PCA and feature selection.
Association rule mining is done using the titanic dataset and a proper case study
analysis is presented in this chapter.
Chapter 7 is mainly to explore the various essential case studies such as text
analytics, credit risk analysis, social network analysis and few exploratory data
analysis. The main purpose of this chapter is to use the basic and advanced concepts
presented in the other previous chapters of this book.

ACKNOWLEDGEMENTS
One of the authors (Dr. G. Sudhamathy) thanks the authorities of Avinashilingam

University, Coimbatore, for providing all the support for making this book a reality.
The author expresses her reverential gratitude to Shri. Dr. P. R. Krishnakumar,

Chancellor, Dr. Premavathy Vijayan, Vice Chancellor and Dr. A. Kowsalya, Registrar,
Avinashilingam Universty, Coimbatore, for providing the opportunity to publish
this book.
The author would like to mention her special regards and thanks to Dr. G.
P. Jeyanthi, Research and Consultancy Director, Dr. A. Parvathi, Dean, Faculty of
Science and Dr. V. Radha, Head, Department of Computer Science, Avinashilingam
Universty, Coimbatore, for their constant encouragement and support to turn this
work into a useful product.
A special thanks to Dr. G. Padmavathi, Professor, Department of Computer

Science, Avinashilingam University, Coimbatore, who was the motivational support
for acquiring the technical knowledge behind this book.
The author wishes to thank all the faculty members of the Department of
Computer Science, Avinashilingam University, Coimbatore, for their continuous
support and suggestions for this book.
We are grateful to the students and teacher community who kept us on our
toes with their constant bombardment of queries which prompted us to learn more,
simplify our learning and findings and place them neatly in a book.
Our Special regards for the experts Mr. Sajeev Madhavan, Director of
Architecture, Oracle, USA and Dr. S. Justus, Associate Professor, VIT, Chennai who
gave their expert opinion in shaping this book into a more appealing format.

xiv Acknowledgement
Most importantly we would like to thank our family members without whose
support this book would not have been a reality.
Last, but not the least, this work is a dedication to God, the Almighty whose
grace has showered upon us in making our dream come true.
G. Sudhamathy
C. Jothi Venkateswaran

Contents
Chapter 1 Basics of R 1
Chapter 2 Data Types in R 27
Chapter 3 Data Preparation 83
Chapter 4 Graphics using R 117
Chapter 5 Statistical Analysis Using R 141
Chapter 6 Data Mining Using R 177
Chapter 7 Case Studies 233
Glossary 299
Packages Used 309
Functions Used 313
References 359
Books 359
Websites 359
Index 361

CHAPTER 1
Basics of R
 OBJECTIVES
On completion of this Chapter you will be able to:

understand how R is different from other languages
install R in your system
write a beginners program using R
get help in R
assign variables in R
know the basic mathematical operations in R
understand various environments and scope of variables
understand functions in R
understand program flow control in R
understand loops in R
1.1. Introducing R
R is a Programming Language and R also refers to the software that is used to run
the R programs. Ross Ihaka and Robert Gentleman from University of Auckland
created R language in 1990s. R language is based on the S language. S Language
was developed at the Bell Laboratories in 1970s. S Language was developed by John
Chambers. R Software is a GNU project free and open source software. R (Language
and Software) is developed by the R Core Team. R has evolved over the past 3 to 4
decades as its history originated from 1970s.
2 R Programming — An Approach for Data Analytics
One can write a new package in R if the existing package is not sufficient
for the individual’s use. R is a high-level scripting language which need not be
compiled, but it is an interpreted language. R is an imperative language and still
it supports object-oriented programming.
R is a free open source language that has cross platform compatibility. R is a

most advanced statistical programming language and it can produce outstanding
graphical outputs. R is extremely flexible and comprehensive even for the beginners.
R easily relates to other programming languages such as C, C++, Java, Python,
Hadoop, etc. R can handle huge data in flat files even in semi structured or in
unstructured form.
The R language allows the user to program loops to successively analyze several
data sets. It is also possible to combine in single program different statistical
functions to perform more complex analyses. The R users may get benefitted from
a large number of programs written and available on the internet. At first R can
look very complex for a beginner or non-specialist. But, this is not actually true as
the prominent feature of R is its flexibility. R displays the results of the analysis
immediately and these results are stored in “objects” so that further analysis can be
done on them. The user can also extract a part of the result which is of interest to
him.
Looking at the features of R, some users may think that “I can’t write programs
using R”. But, this is not the case for two reasons. First, R is an interpreted language
and not a compiled one. This means that all commands typed on the keyboard are
directly executed without need to build the complete program like in C, C++ or
Java. Second, R’s syntax is very simple and intuitive.
In R, a function is always written with parentheses, eg. ls(). If only the name
of the function is typed, R displays the content of the function. In this book the
functions are written with their names followed by parentheses to distinguish them
from other objects. When R is running variables, data, functions, results, etc. are
stored in the active memory of the computer in the form of objects which have a
name. The user can do actions on these objects with operators and functions.
3 Basics of R
1.2. Installing R
R is available in several forms, essentially for Unix and Linux machines, or some
pre-compiled binaries for Windows, Linux and Macintosh. The files needed to
install R, either from the source or from the pre-compiled binaries are distributed
from the internet site of the Comprehensive R Archive Network (CRAN) where the
instructions for installation are also available.
R can be installed from the link http://www.r-project.org using internet

connection. Use the “Download R” link in web page to download the R Executable.
Choose the version of R that is suitable for your operating system. R-Scripts can
run without the installation of the IDE, the R-Studio using the R-Console. The
prerequisite for installing R-Studio is that one should have downloaded and installed
any version of R. (Version 3.3.0 of R is used for installation and running scripts used
in this book). Follow the instructions on the website to complete installation of R
Console.
Once R installation is completed we install R-Studio. For installation of

R-Studio in Windows operating system, we download the latest precompiled
binary distribution from the CRAN website http://www.rstudio.org. (Version 3.4 of
R-Studio is used for installation and running scripts used in this book). Start the
installation and follow the steps required by the setup wizard. Once completed,
launch RStudio IDE from Start à All Programs à Rstudio à RStudio.exe or from your
custom installation directory. The default installation directory for RStudio IDE is
“C:\Program Files\RStudio\bin\rstudio.exe.
R Studio is an Integrated Development Environment (IDE) that consists of

a GUI with four parts – 1) A text editor 2) command-line interpreter 3) place to
display files, plots, packages and help information 4) place to display the data being
used and the variables used in the program (Environment / History).
Figure 1.1 R-Studio GUI
1.3. Initiating R
1.3.1. First Program
Open R Gui, find the command prompt and type the command below and hit enter
to run the command.
> sum(1:5)
[1] 15
The result above shows that the command gives the result 15. That is the
command has taken the input of integers from 1 to 5 and has performed the sum
operation on them. In the above command sum() is a function that takes the
argument 1:5 which means a vector that consists of a sequence of integers from 1 to
5. Like any other command prompt, R also allows to use the up arrow key to revoke
the previous commands.
1.3.2. Help in R
There are many ways to get help from R. If a function name or a dataset name is
known then we can type ? followed by the name. If name is not known then we
5 Basics of R
need to type ?? followed by a term that is related to the search function. Keywords,
special characters and two separate terms of search need to be enclosed in double or
single quotes. The symbol # is used to comment a line in R Program like any other
programming language.
> ?mean # help page for mean function opens
> ?”+” # help page for addition function opens
> ?”if ” # help page for if opens
> ??plotting # searches for the help pages containing the word “plotting”
> ??”regression model” # searches for “regression model” phrase
The same help can be obtained by the functions help() and help.search(). In
these functions the arguments has to be enclosed by quotes.
> help(“mean”)
> help(“+”)
> help(“if ”)
> help.search(“plotting”)
> help.search(“regression model”)
1.3.3. Assigning Variables

The results of the operations in R can be stored for reuse. The values can be assigned
to the variables using the symbol “<-” or “=” of which the symbol “<-” is preferred.
There is no concept of variables declaration in R. The variable type is assumed
based on the value assigned.
> X <- 1:3
>X
[1] 1 2 3
> Y = 4:6
>Y
[1] 4 5 6
>X+3*Y-2
[1] 11 15 19
The variable names consist of letters, numbers, dots and underscores, but a
variable name should only start with an alphabet. The variable names should not
be reserve words. To create global variables (variables available everywhere) we use
the symbol “<<-”.
X <<- exp(exp(1))
Assignment operation can also be done using the assign() function. For global
assignment the same function assign() can be used, but, by including an extra
attribute globalenv(). To see the value of the variable, simply type the variable in
the command prompt. The same thing can be done using a print() function.
> assign(“F”, 3 * 8)
> assign(“G”, 6 * 9, globalenv())
>F
[1] 24
> print(G)
[1] 54
If assignment and printing of a value has to be done in one line we can do the
same in two ways. First method, by separating the two statements by a semicolon
and the second method is by wrapping the assignment in parenthesis () as below.
> L <- sum(4:8); L
[1] 30
> (M <- sum(5:9))
[1] 35
1.3.4. Basic Mathematical Operations
The “+” plus operator is used to perform the addition operation. It can be used
to add two numbers or add two vectors. Vector represents an ordered set of values.
Vectors are mainly used to analyse statistical data. The “:” colon operator creates
a sequence. Sequence is a series of numbers within the given limits. The “c()”
function concatenates the values given within the brackets “(“ and “)”. Variable
7 Basics of R
names in R are case sensitive. Open R Gui, find the command prompt and type the
command below and hit enter to run the command.
> 7:12 + 12:17
[1] 19 21 23 25 27 29
> c(3, 1, 8, 6, 7) + c(9, 2, 5, 7, 1)
[1] 12 3 13 13 8
The vectors and the c() function in R helps us to avoid loops. The statistical
functions in R can take the vectors as input and produce results. The sum() function
takes vector arguments and produces results. But, the median() function when
taking the vector arguments shows errors.
> sum(7:10)
[1] 34
> mean(7:10)
[1] 8.5
> median(7:10)
[1] 8.5
> sum(7,8,9,10)
[1] 34
> mean(7,8,9,10)
[1] 7
> median(7,8,9,10)
Error in median(7, 8, 9, 10) : unused arguments (9, 10)
Similar to the “+” plus operator all other operators in R take vectors as inputs
and can produce results. The subtraction and the multiplication operations work
as below.
> c(5, 6, 1, 9) - 2
[1] 3 4 -1 7
> c(5, 6, 1, 9) - c(4, 2, 0, 7)
[1] 1 4 1 2
> -1:4 * -2:3
[1] 2 0 0 2 6 12
> -1:4 * 3
[1] -3 0 3 6 9 12
The exponentiation operator is represented using the symbol “^” or the “**”.
This can be checked using the function identical().
> identical(2^3, 2**3)
[1] TRUE
The division operator is of three types. The ordinary division is represented

using the “/” symbol, the integer division operator is represented using the “%/%”
symbol and the modulo division operator is represented using the “%%” symbol.
The below example commands show the results of the division operators.
> 5:9/2
[1] 2.5 3.0 3.5 4.0 4.5
> 5:9%/%2
[1] 2 3 3 4 4
> 5:9%%2
[1] 1 0 1 0 1
The other mathematical functions are the trigonometry functions like, sin(),
cos(), tan(), asin(), acos(), atan() and the logarithmic and exponential functions
like log(), exp(), log1p(), expm1(). All these mathematical functions can operate on
vectors as well as individual elements. Few more examples of the mathematical
functions are listed below
The operator “==” is used for comparing two values. For checking inequalities
of values the operator “!=” is used. These operators are called the relational
operators. The relational operators also take the vectors as input and operate on
them. The other relational operators are the “< “, “> “, “<= “ and “>= “.
> c(2, 4 - 2, 1 + 1) == 2
9 Basics of R
[1] TRUE TRUE TRUE

> 1:5 != 5:1
[1] TRUE TRUE FALSE TRUE TRUE
> exp(1:3) < 20
[1] TRUE TRUE FALSE
> (1:10) ^ 2 >= 50
[1] FALSE FALSE FALSE FALSE FALSE FALSE FALSE TRUE TRUE TRUE
Non-integers cannot be compared using the operator “==” as it produces

wrong results due to rounding off error of the float numbers being compared. For
overcoming this issue we have the function all.equal(). If the value to be compared
by the function all.equal() is not equal, it returns a report on the difference. To
get a TRUE or FALSE reply, the all.equal() function has to be wrapped using the
function isTRUE(). The below example code will help to understand the concepts
discussed.
> sqrt(2) ^ 2 == 2
[1] FALSE
> sqrt(2) ^ 2 - 2
[1] 4.440892e-16
> all.equal(sqrt(2) ^ 2, 2)
[1] TRUE
> all.equal(sqrt(2) ^ 2, 3)
[1] “Mean relative difference: 0.5”
> isTRUE(all.equal(sqrt(2) ^ 2, 3))
[1] FALSE
The equality operator “==” can also be used to compare strings, but, string
comparison is case sensitive. Similarly, the operators “<” and “>” can also be used
on strings. The below examples show the results.
> c(“Week”, “WEEK”, “week”, “weak”) == “week”
[1] FALSE FALSE TRUE FALSE
> c(“A”, “B”, “C”) < “B”

[1] TRUE FALSE FALSE
> c(“a”, “b”, “c”) < “B”
[1] TRUE TRUE FALSE
1.4. Packages in R
R Packages are installed in an online repository called CRAN (Comprehensive R
Archive Network). A Package is a collection of R functions and datasets. Currently,
the CRAN package repository features 10756 available packages. The list of all
available packages in the CRAN repository can be viewed from the web site “https://
cran.r-project.org/web/packages/available_packages_by_name.html”. To find the
list of functions available in a package (say the package is “stats”) we can use the
command ls(“package:stats”) or the command library(help = stats) in the command
prompt.
A library is a folder in the machine that stores the files for a package. If a package
is already installed on a machine we can load the same using the library() function.
The name of the package to be loaded is passed to the library() function as argument
without enclosing in quotes. If the package name has to be programmatically passed
to the library() function, then we need to set the argument character.only = TRUE.
If a package is not installed and if the library() function is used to load the package,
it will throw an error message. Alternatively if the require() function is used to load
a package, it returns TRUE if the package is already installed or it returns FALSE if
the package is not already installed.
We can list and see all the packages that are already loaded using the search()
function. This list shows the global environment as the first one followed by the
recently loaded packages. The last two are special environments, namely, “Autoloads”
and “base” package.
> search()
[1] “.GlobalEnv” “package:cluster” “tools:rstudio” “package:stats”
[5] “package:graphics” “package:grDevices” “package:utils” “package:datasets”
[9] “package:methods” “Autoloads” “package:base”
11 Basics of R
The function installed.packages() returns a data frame with information about

all the packages installed in a machine. It is safe to view the results of this using
the View() function as it may list hundreds of packages. This list of packages also
shows the version of the package installed, location on the machine and dependent
packages.
> View(installed.packages())
The function R.home(“library”) retrieves the location on the machine that

stores all R default packages. The same result can be accomplished using the .Library
command. The home directory can be listed using the path.expand(“~”) and Sys.
getenv(“HOME”) functions.
> R.home(“library”)
[1] “C:/PROGRA~1/R/R-33~1.0/library”
> .Library
[1] “C:/PROGRA~1/R/R-33~1.0/library”
> path.expand(“~”)
[1] “C:/Users/admin/Documents”
> Sys.getenv(“HOME”)
When R is upgraded, it is required to reinstall all the packages as different

versions of R needs different versions of the packages. The function .libPaths() lists
all the R libraries in the installed machine. The first value listed is the place where
the packages will be installed by default.
> .libPaths()
[1] “C:/Users/admin/Documents/R/win-library/3.3”
[2] “C:/Program Files/R/R-3.3.0/library”
The CRAN package repository contains handful of packages that needs special
attention. To access additional repositories, type setRepositories() and select
the repository required. The repositories R-Forge and rforge.net contains the
development versions of the packages that appear on the CRAN repository. The
function available.packages() lists thousands of packages in each of the selected
repository. (Note: can use the View() function to restrict fetching of thousands of
the packages at one go)
> setRepositories()
--- Please select repositories for use in this session ---
1: + CRAN
2: BioC software
3: BioC annotation
4: BioC experiment
5: BioC extra
6: CRAN (extras)
7: Omegahat
8: R-Forge
9: rforge.net
10: + CRANextra
Enter one or more numbers separated by spaces, or an empty line to cancel
1:
There are many online repositories like GitHub, Bitbucket, and Google Code
from where many R Packages can be retrieved. The packages can be installed using
the function install.packages() function by mentioning the name of the package as
argument to this function. But, it is necessary to have internet connection to install
any package and write permission to the hard drive. To update the latest version of
the installed packages, we use the function update.packages() with the argument
ask = FALSE which disallows prompting before updating each package. To delete
a package already installed, we use the function remove.packages() by passing the
name of the package to be removed as argument.
> install.packages(“chron”)
13 Basics of R
1.5. Environments and Functions
1.5.1. Environments
In R the variables that we create need to be stored in an environment. Environments

are another type of variables. We can assign them, manipulate them and pass
them as arguments to functions. They are like lists that are used to store different
types of variables. When a variable is assigned in the command prompt, it goes by
default into the global environment. When a function is called, an environment is
automatically created to store the function-related variables. A new environment is
created using the function new.env().
> newenvironment <- new.env()
We can assign variables into a newly created environment using the double
square brackets or the dollar operator as below.
> newenvironment[[“variable1”]] <- c(4, 7, 9)
> newenvironment$variable2 <- TRUE
> assign(“variable3”, “Value for variable3”, newenvironment)
The assign() function can also be used to assign variables to an environment.

Retrieving values stored in an environment is like list indexing or we can use the
get() function.
> newenvironment[[“variable1”]]
[1] 4 7 9
> newenvironment$variable2
[1] TRUE
> get(“variable3”, newenvironment)
[1] “Value for variable3”
The functions ls() and ls.str() take an environment argument and lists its
contents. We can test if a variable exists in an environment using the exists()
function.
> ls(envir = newenvironment)

[1] “variable1” “variable2” “variable3”
> ls.str(envir = newenvironment)
variable1 : num [1:3] 4 7 9
variable2 : logi TRUE
variable3 : chr “Value for variable3”
> exists(“variable2”, newenvironment)
[1] TRUE
An environment can be converted into a list using the function as.list() and a
list can be converted into an environment using the function as.environment() or
the function list2env().
> newlist <- as.list(newenvironment)
> newlist
$variable3
$variable1
[1] 4 7 9
$variable2
[1] TRUE
> as.environment(newlist)
<environment: 0x124730a8>
> list2env(newlist)
<environment: 0x12edf3e8>
> anotherenv <- as.environment(newlist)
> anotherenv[[“variable3”]]
All environments are nested and so every environment has a parent environment.
The empty environment sits at the top of the hierarchy without any parent. The
15 Basics of R
exists() and the get() function also looks for the variables in the parent environment.
To change this behaviour we need to pass the argument inherits = FALSE.
> subenv <- new.env(parent = newenvironment)
> exists(“variable1”, subenv)
[1] TRUE
> exists(“variable1”, subenv, inherits = FALSE)
[1] FALSE
The word frame is used interchangeably with the word environment. The function
to refer to parent environment is denoted as parent.frame(). The variables assigned
from the command prompt are stored in the global environment. The functions and
the variables from the R’s base package are stored in the base environment.
1.5.2. Functions
A function and its environment together is called a closure. When we load a
package, the functions in that package are stored in the environment on the search
path where the package is installed. A function in R is a verb and not a noun as it
does things with its data. Functions are also another data types and hence we can
assign and manipulate and pass them as arguments to other functions. Typing the
function name in the command prompt lists the code associated with the function.
Below is the code listed for the function readLines().
> readLines
function (con = stdin(), n = -1L, ok = TRUE, warn = TRUE, encoding = unknown”,
skipNul = FALSE)
{
if (is.character(con)) {
con <- file(con, “r”)
on.exit(close(con))
}
.Internal(readLines(con, n, ok, warn, encoding, skipNul))
}
When we call a function by passing values to it, the values are called as
arguments. The lines of code of the function can be seen between the curly braces
as body of the function. In R, there is no explicit return statement to return values.
The last value that is calculated in a function is returned by default in R.
To create user defined functions, it is required to just assign the function as we

do for other variables. The below code is an example of how to create a user defined
function. In this cube is the name of the function and x is the argument passed
to this function. The content within the curly braces is the body of the function.
(Note: If it is a one line code we can omit the curly braces). Once a function is
defined, it can be called like any other function in R by passing its arguments.
> cube <- function(x)
{ cu <- x ^ 3}
> z <- cube(5)
>z
[1] 125
The functions formals(), args() and formalArgs() can fetch the arguments
defined for a function. The body of the function can be retrieved using the body()
and deparse() functions.
> formals(cube)
$x
> args(cube)
function (x)
NULL
> formalArgs(cube)
[1] “x”
> body(cube)
{
cu <- x^3
}
17 Basics of R
> deparse(cube)
[1] “function (x) “ “{“ “ cu <- x^3” “}”
Functions can be passed as arguments to other functions and they can be

returned from other functions. For calling a function, there is another function
called do.call() in which we can pass the function name and its arguments as
arguments. The use of this function can be seen below when using the rbind()
function to concatenate two data frames.
> f1 <- data.frame(x = 1:4, y = 5:8)
> f2 <- data.frame(x = 9:12, y = 13:16)
> do.call(rbind, list(f1, f2))
x y
1 1 5
2 2 6
3 3 7
4 4 8
5 9 13
6 10 14
7 11 15
8 12 16
When using functions as arguments to the do.call() function, it is not necessary

to assign them first. We can pass a function anonymously as below.
> do.call(function(x,y) x * y, list(1:3, 4:6))
[1] 4 10 18
1.5.3. Variable Scope

Variable’s scope is the place where we can see the variable. If a variable is defined
within a function, the variable can be accessed from any statement in the function.
Also the sub-functions will have access to the variables defined in the parent
function.
> x <- function(a1)

{
a2 <- 1
y <- function(a1)
{
a2 / a1
}
y(a1)
}
> x(5)
[1] 0.2
Thus R will search for a variable in the current environment and if it could not
find it, it will check the same in its parent environment. This search will proceed
upwards until the variable is searched in the global environment. The variables
defined in the global environment are called the global variables, which can be
accessed from anywhere else. The replicate() function can be used to run a function
several times as below. In this the user defined function random() returns 1 if the
value returned by the rnorm() function is a positive value and otherwise it returns
the value of the argument passed to the function random(). This function random()
is called 20 times using the replicate() function.
> random <- function(x)
+{
+ if(rnorm(1) > 0)
+ {r <- 1}
+ else
+ {r <- x}
+}
> replicate(20, random(5))
[1] 5 5 1 1 5 1 5 5 5 5 5 5 5 5 5 1 1 5 1 5
19 Basics of R
1.6. Flow Control

In some situations it may be required to execute some code only if a condition is
satisfied.
1.6.1. If and Else Statement
The if statement takes a logical value and executes the next statement only if the
value is TRUE.
> if(TRUE) message(“TRUE Statement”)
TRUE Statement
> if(FALSE) message(“FALSE Statement”)
It is not necessary to pass the logical values as TRUE or FALSE directly, instead
a variable or expression that returns a logical value can be used. If there are several
statements to execute after the condition, they can be wrapped in curly braces.
a <- 5
if(a < 7)
{
b <- a * 5
c <- b * 3
message(“b is “, b)
message(“c is “, c)
}
b is 25
c is 75
In the if and else construct the code that follows the if statement is executed if
the condition is TRUE and the code that follows the else statement is executed if
the condition is FALSE. It is important to note that the else statement must occur
on the same line as the closing curly brace of the if statement and otherwise it will
throw an error message.
a <- 8
if(a < 7)
{
b <- a * 5
c <- b * 3
message(“b is “, b)
message(“c is “, c)
} else
{
message(“a is greater than 7”)
}
a is greater than 7
The if and else statements can be used repeatedly to code multiple conditions
and this respective actions. In this case it is important to note that the if and the
else statements are separated and they are not one word as ifelse. The ifelse function
is of different use which will be covered shortly.
a <- -8
if(a < 0)
{
message(“a is negative”)
} else if(a == 0)
{
message(“a is zero”)
} else if(a > 0)
{
message(“a is positive”)
a is negative
21 Basics of R
The ifelse() function takes three arguments of which the first is logical condition,
the second is the value that is returned when the first vector is TRUE and the third
is the value that is returned when the first vector is FALSE.
> a <- 3
> b <- 5
> ifelse(a < b, “a is less than b”, “a is greater than b”)
[1] “a is less than b”
1.6.2. Switch Statement
If there are many else statements, it looks confusing and in such cases the switch()
function is required. The first argument of the switch statement is an expression that
can return a string value or an integer. This is followed by several named arguments
that provide the results when the name matches the value of the first argument.
Here also we can execute multiple statements enclosed by curly braces. If there is
no match the switch statement returns NULL. So, in this case, it is safe to mention
a default value if none matches.
> switch(“color”,”color” = “red”, “shape” = “circle”, “radius” = 10)
[1] “red”
> switch(“position”,”color” = “red”, “shape” = “circle”, “radius” = 10)
[1] NULL
> switch(“position”,”color” = “red”, “shape” = “circle”, “radius” = 10,”default”)
[1] “default”
> switch(2,”red”,”green”,”blue”)
[1] “green”
1.7. Loops
There are three kinds of loops in R namely, repeat, while and for.
1.7.1. Repeat Loops
The repeat is the easiest loop in R that executes the same code until it is forced to
stop. This repeat is similar to the do while statement in other languages. A break
statement can be given when it is required to break the looping. Also, it is possible
to skip the rest of the statements in a loop and execute the next iteration and this
is done using the next statement.
a <- 1
repeat {
message(“Inside the loop”)
if(a == 3)
{
a=a+1
next
}
message(“The value of a is “, a)
a=a+1
if(a > 5)
{
message(“Exiting the loop”)
break
}
}
Inside the loop
The value of a is 1
Inside the loop
The value of a is 2
Inside the loop
Inside the loop
23 Basics of R
The value of a is 4
Inside the loop
The value of a is 5
Exiting the loop
1.7.2. While Loops
The while loops are backward repeat loops. The repeat loop executes the code and
then checks for the condition, but in while loops the condition is first checked
and then the code is executed. So, in this case it is possible that the code may not
be executed even once when the condition fails at the entry itself during the first
iteration. The same example above can be written using the while statement.
a <- 1
while (a <= 5)
{
message(“Inside the loop”)
if(a == 3)
{
a=a+1
next
}
message(“The value of a is “, a)
a=a+1
}
Inside the loop
The value of a is 1
Inside the loop
The value of a is 2
Inside the loop
Inside the loop
The value of a is 4
Inside the loop
The value of a is 5
1.7.3. For Loops
The for loops are used when we know how many times the code needs to be repeated.
The for loop accepts an iterating variable and a vector. It repeats the loop giving the
iterating each element from the vector in turn. In this case also if there are multiple
statements to execute, we can use the curly braces. The iterating variable can be an
integer, number, character or logical vectors and they can be even lists.
for(i in 1:5)
{
j <- i * i
message(“The square value of “, i, “ is “, j)
}
The square value of 1 is 1
for(i in c(TRUE, FALSE, NA))
{
message(“This Statement is “, i)
}
This Statement is TRUE
This Statement is FALSE
This Statement is NA
a <- c(1,2,3)
b <- c(“a”,”b”,”c”,”d”)
25 Basics of R
d <- c(TRUE, FALSE)

l = list(a, b, d)
for(i in l)
{
message(“The value of the list is “, i)
}
The value of the list is 123
The value of the list is abcd
The value of the list is TRUEFALSE
 HIGHLIGHTS
 R is a free open source language that has cross platform compatibility.
 R’s syntax is very simple and intuitive.
 R’s installation software can be downloaded from the CRAN Website.
 Help in R can be obtained by using, for eg. ?mean() / help(“mean”)
 Variables can be assigned using the symbol ß or the assign() function.
 The basic functions are c(), sum(), mean(), median(), exp(), sqrt() etc.
 The basic operators are “+”, “*”, “:”, “/”, “**”, “*”, “%%”, “%/%”, “==”,
“!=”, “<”, “>”, “<=”, “>=” etc.
 Currently, the CRAN package repository features 10756 available packages.
 A Package can be newly installed using the function install.packages() and
it can be invoked using the function library().
 When a variable is assigned in the command prompt, it goes by default
into the global environment.
 To create a new environment we use the function new.env().
 Typing the function name in the command prompt lists the code
associated with the function.
 The if and the else statements are separated and they are not one word as
ifelse.
 The ifelse() function takes three arguments.
 If there are many else statements, the switch() function is required.
 The function repeat is similar to the do while statement in other languages.

 A break statement can be given when it is required to break the looping.
 To skip the rest of the statements in a loop we use the next statement.
 The while loops are backward repeat loops.
 The for loops are used when we know how many times the code needs to
be repeated.
CHAPTER 2
Data Types in R
 OBJECTIVES

know the basic data types in R of which the other complex data types are
made of
know how to create, access and perform basic operations on the vector
data types in R
know how to create, access and perform basic operations on matrices and
arrays in R
know how to create, access and perform basic operations on the list data
types
know how to create, access and perform basic operations on the factor
data types in R
know how to create, access and perform basic operations on strings in R
understand the various date and time classes in R
convert between various date formats
setup various time zones
perform calculations on dates and times
2.1. Basic Data Types in R

In contrast to other programming languages like C and Java, in R, the variables are
not declared as some data type. The variables are assigned with R-Objects and the
data type of the R-objects becomes the data type of the variables. There are many
types of R-objects. The frequently used ones are − Vectors, Arrays, Matrices, Lists,
Data Frames, Strings and Factors.
The simplest of these objects is the Vector object and there are six data types
of these atomic vectors, also termed as six classes of vectors. The other R-Objects
are built upon the atomic vectors. Hence, the basic data types in R are Numeric,
Integer, Complex, Logical and Character.
2.1.1. Numeric
Decimal values are called numeric in R. It is the default computational data type. If
we assign a decimal value to a variable x as follows, x will be of numeric type.
> x = 10.5
> x
[1] 10.5
> class(x) # print the class name of x
[1] “numeric”
Further more, even if we assign an integer to a variable k, it is still being saved as

a numeric value. The fact that if k is an integer can be confirmed with the is.integer()
function.
>k=1
>k
[1] 1
> class(k)
[1] “numeric”
> is.integer(k)
[1] FALSE
2.1.2. Integer
In order to create an integer variable in R, the as.integer() function is invoked as

below.
29 Data Types in R
> y = as.integer(3)
>y
[1] 3
> class(y)
[1] “integer”
> is.integer(y)
[1] TRUE
We can force a numeric value into an integer with the same as.integer() function
as below.
> as.integer(3.14)
[1] 3
Similarly we can parse a string for a decimal value as below.

> as.integer(“5.27”) # force a decimal string
[1] 5
But, if a non decimal string is forced, it is an error and it returns NA.

> as.integer(“abc”)
[1] NA
Warning message:
NAs introduced by coercion
The integer values of the logical values TRUE and FALSE are 1 and 0 respectively.
> as.integer(TRUE)
[1] 1
> as.integer(FALSE)
[1] 0
2.1.3. Complex
A complex number is expressed as an imaginary value i.

> z = 3 + 4i
>z
[1] 3 + 4i
> class(z)
[1] “complex”
If we find the square root of -1, it gives an error. But, if it is converted into a
complex number and then square root is applied, it produces the necessary result as
another complex number.
> sqrt(-1)
[1] NaN
Warning message:
In sqrt(-1) : NaNs produced
> sqrt(as.complex(-1))
[1] 0+1i
2.1.4. Logical
When two variables are compared, the logical values are created. The logical
operators are “&” (and), “|” (or), and “!” (negation).
> a = 4; b = 7
>p=a>b
>p
[1] FALSE
> class(p)
[1] “logical”
> a = TRUE; b = FALSE
>a&b
[1] FALSE
>a|b
31 Data Types in R
[1] TRUE
> !a
[1] FALSE
2.1.5. Character
The character object is used to represent string values in R. Objects can be converted
into character values using the as.character() function. A paste() function can be
used to concatenate two character values.
> s = as.character(“7.48”)
>s
[1] “7.48”
> class(s)
[1] “character”
> fname = “Adam”
> lname = “Smith”
> paste(fname, lname)
[1] “Adam Smith”
However, a readable string can be created using the sprint() function and this is
similar to the C language syntax.
> sprintf(“%s has %d rupees”, “Sundar”,1000)
[1] “Sundar has 1000 rupees”
The substr() function can be used to extract a substring from a given string. The
sub() function is used to replace the first occurrence of a string with another string
as below.
> substr(“Twinkle Twinkle Little Star”, start = 9, stop = 15)
[1] “Twinkle”
> sub(“Twinkle”, “Wrinkle”, “Twinkle Twinkle Little Star”)
[1] “Wrinkle Twinkle Little Star”
2.2. Vectors
A sequence of data elements of the same basic type is called a Vector. Members in a
vector are called as components or members. The vector() function creates a vector
of a specified type and length. The result is a zero or FALSE or empty string.
> vector(“numeric”, 3)
[1] 0 0 0
> vector(“logical”, 5)
[1] FALSE FALSE FALSE FALSE FALSE
> vector(“character”, 2)
[1] “” “”
The below commands also produces the same result as the above commands.
> numeric(3)
[1] 0 0 0
> logical(5)
[1] FALSE FALSE FALSE FALSE FALSE
> character(2)
[1] “” “”
The seq() function allows to generate sequences. The function seq.int() also
creates sequence from one number to another, but this function provides more
options for splitting the sequence.
> seq(1:5)
[1] 1 2 3 4 5
> seq.int(5, 12)
[1] 5 6 7 8 9 10 11 12
> seq.int(10, 5, -1.5)
[1] 10.0 8.5 7.0 5.5
33 Data Types in R
The function seq_len() creates a sequence from 1 to the input value. The
function seq_along() creates a sequence from 1 to the length of the input.
> seq_len(7)
[1] 1 2 3 4 5 6 7
> p <- c(3, 4, 5, 6)
> seq_along(p)
[1] 1 2 3 4
The function length() can be used to find the length of the vector, that is the
number of elements in a vector. Using this function, it is possible to assign new
length to a vector. If the vector length is extended NA(s) will be added to the end.
> length(1:7)
[1] 7
> length(c(“aa”, “ccc”, “eeee”))
[1] 3
> nchar(c(“aa”, “ccc”, “eeee”))
[1] 2 3 4
> s <- c(1,2,3,4,5)
> length(s) <- 3
>s
[1] 1 2 3
> length(s) <- 8
>s
[1] 1 2 3 NA NA NA NA NA
Each element of a vector can be given a name during the vector creation itself. If
there are space or special characters in the name, it needs to be enclosed in quotes.
The names() function can be used to give names to the vector elements after its
creation.
> c(a = 1, b = 2, c = 3)
abc
123
> s <- 1:3
>s
[1] 1 2 3
> names(s) <- c(“a”, “b”, “c”)
>s
abc
123
Elements of a vector can be accessed using its indexes which are specified in
a square bracket. The index number starts from 1 and not 0. Specifying a negative
number as index to a vector means, it returns all the elements except the one
specified. The name of the vector element can also be specified as index to fetch it.
> x <- c(1:5)
>x
[1] 1 2 3 4 5
> x[c(2,3)]
[1] 2 3
> x[c(-1,-4)]
[1] 2 3 5
> s <- 1:3
>s
[1] 1 2 3
> names(s) <- c(“a”, “b”, “c”)
> s[“b”]
b
2
35 Data Types in R
If an incorrect index is specified to access a vector element, the result is NA.

Non integer indices are rounded off. Not passing any index to a vector will return all
the elements of the vector.
>x
[1] 1 2 3 4 5
> x[7]
[1] NA
The which() function returns the elements of the vector which satisfies the
condition specified within this function. The functions which.min() and which.
max() can be used to display the minimum and the maximum elements in the
vector.
>x
[1] 1 2 3 4 5
> which.min(x)
[1] 1
> which.max(x)
[1] 5
> which(x>3)
[1] 4 5
Vectors can be combined using the c() function. When the two vectors are
combined the numeric values are forced into character values. This shows that all
the members of a vector should be of the same basic data type.
> f = c(7, 5, 9)
> g = c(“aaa”, “bbb”, “ccc”)
> c(f, g)
[1] “7” “5” “9” “aaa” “bbb” “ccc”
Arithmetic operations in a vector will be performed member-wise. If two vectors

are of unequal length, the shorter vector will be recycled in order to match the
longer vector.
> x = c(5, 8, 9)
> y = c(2, 6, 9)
>4*y
[1] 8 24 36
>x+y
[1] 7 14 18
>x-y
[1] 3 2 0
>x*y
[1] 10 48 81
>x/y
[1] 2.500000 1.333333 1.000000
> v = c(1, 2, 3, 4, 5, 6)
>x+v
[1] 6 10 12 9 13 15
The rep() function creates a vector with repeated elements. This function has
its other variants such as rep.int() and rep_len() whose usage is as given below.
> rep(1:3, 4)
[1] 1 2 3 1 2 3 1 2 3 1 2 3
> rep(1:3, each = 4)
[1] 1 1 1 1 2 2 2 2 3 3 3 3
> rep(1:3, times = 1:3)
[1] 1 2 2 3 3 3
> rep(1:3, length.out = 9)
[1] 1 2 3 1 2 3 1 2 3
> rep.int(1:3, 4)
[1] 1 2 3 1 2 3 1 2 3 1 2 3
37 Data Types in R
> rep_len(1:3, 9)
[1] 1 2 3 1 2 3 1 2 3
2.3. Matrices and Arrays

A matrix is a collection of data elements with the same basic type arranged in a two-
dimensional rectangular layout. An array consists of multidimensional rectangular
data. Matrices are special cases of two-dimensional arrays. To create an array the
array() function can be used and a vector of values and vector of dimensions are
passed to it.
> x <- array(1:24, dim = c(4, 3, 2),
dimnames = list(c(“a”, “b”, “c”, “d”), c(“e”, “f ”, “g”), c(“h”, “i”)))
>x
,,h
ef g
a15 9
b 2 6 10
c 3 7 11
d 4 8 12
,,i
e f g
a 13 17 21
b 14 18 22
c 15 19 23
d 16 20 24
The syntax for creating matrices is using the function matrix() and passing the
nrow or ncol argument instead of the dim argument in the arrays. A matrix can also
be created using the array() function where the dimension of the array is two.
> m <- matrix(1:12, nrow = 3, dimnames = list(c(“a”, “b”, “c”), c(“d”, “e”, “f ”, “g”)))
>m
def g
a 1 4 7 10
b 2 5 8 11
c 3 6 9 12
> m1 <- array(1:12, dim = c(3,4),
dimnames = list(c(“a”, “b”, “c”), c(“d”, “e”, “f ”, “g”)))
> m1
def g
a 1 4 7 10
b 2 5 8 11
c 3 6 9 12
The argument byrow = TRUE in the matrix() function assigns the elements
row wise. If this argument is not specified, by default the elements are filled column
wise.
> m <- matrix(1:12, nrow = 3, byrow = TRUE,
dimnames = list(c(“a”, “b”, “c”), c(“d”, “e”, “f ”, “g”)))
The dim() function returns the dimensions of an array or a matrix. The functions
nrow() and ncol() returns the number of rows and number of columns of a matrix
respectively.
> dim(x)
[1] 4 3 2
> dim(m)
[1] 3 4
> nrow(m)
[1] 3
> ncol(m)
[1] 4
39 Data Types in R
The length() function also works for matrices and arrays. It is also possible to
assign new dimension for a matrix or an array using the dim() function.
> length(x)
[1] 24
> length(m)
[1] 12
> dim(m) <- c(6,2)
The functions rownames(), colnames() and dimnames() can be used to fetch the
row names, column names and dimension names of matrices and arrays respectively.
> rownames(m1)
[1] “a” “b” “c”
> colnames(m1)
[1] “d” “e” “f ” “g”
> dimnames(x)
[[1]]
[1] “a” “b” “c” “d”
[[2]]
[1] “e” “f ” “g”
[[3]]
[1] “h” “i”
It is possible to extract the element at the nth row and mth column using the
expression M[n, m]. The entire nth row can be extracted using M[n, ] and similarly,
the mth column can be extracted using M[,m]. Also, it is possible to extract more
than one column or row.
> M[2,3]
[1] 6
> M[2,]
[1] 4 5 6
> M[,3]
[1] 3 6 9
> M[,c(1,3)]
[,1] [,2]
[1,] 1 3
[2,] 4 6
[3,] 7 9
> M[c(1,3),]
[,1] [,2] [,3]
[1,] 1 2 3
[2,] 7 8 9
The matrix transpose is constructed by interchanging its rows and columns

using the function t().
> t(M)
r1 r2 r3
c1 1 4 7
c2 2 5 8
c3 3 6 9
The columns of two matrices can be combined using the cbind() function and
similarly the rows of two matrices can be combined using the rbind() function.
> M1 = matrix(c(2,4,6,8,10,12), nrow=3, ncol=2)
> M1
[,1] [,2]
[1,] 2 8
[2,] 4 10
[3,] 6 12
> M2 = matrix(c(3,6,9), nrow=3, ncol = 1)
> M2
41 Data Types in R
[,1]
[1,] 3
[2,] 6
[3,] 9
> cbind(M1, M2)
[,1] [,2] [,3]
[1,] 2 8 3
[2,] 4 10 6
[3,] 6 12 9
> M3 = matrix(c(4,8), nrow=1, ncol=2)
> M3
[,1] [,2]
[1,] 4 8
> rbind(M1, M3)
[,1] [,2]
[1,] 2 8
[2,] 4 10
[3,] 6 12
[4,] 4 8
A matrix can be deconstructed using the c() function which combines all
column vectors into one.
> c(M1)
[1] 2 4 6 8 10 12
The arithmetic operators “+”, “- “, “* “, “/ “ work element wise on matrices

and arrays. But the condition is that the matrices or arrays should be of conformable
sizes. The matrix multiplication is done using the operator “%*%”.
> M1 = matrix(c(2,4,6,8,10,12), nrow=3, ncol=2)
> M1
[,1] [,2]
[1,] 2 8
[2,] 4 10
[3,] 6 12
> M2 = matrix(c(3,6,9,11,1,5), nrow=3, ncol = 2)
> M2
[,1] [,2]
[1,] 3 11
[2,] 6 1
[3,] 9 5
> M1 + M2
[,1] [,2]
[1,] 5 19
[2,] 10 11
[3,] 15 17
> M1 * M2
[,1] [,2]
[1,] 6 88
[2,] 24 10
[3,] 54 60
> M2 = matrix(c(3,6,9,11), nrow=2, ncol = 2)
> M2
[,1] [,2]
[1,] 3 9
[2,] 6 11
> M1 %*% M2
[,1] [,2]
[1,] 54 106
43 Data Types in R
[2,] 72 146
[3,] 90 186
The power operator “^” also works element wise on matrices. To find the
inverse of a matrix the function solve() can be used.
> M2
[,1] [,2]
[1,] 3 9
[2,] 6 11
> M2^-1
[,1] [,2]
[1,] 0.3333333 0.11111111
[2,] 0.1666667 0.09090909
> solve(M2)
[,1] [,2]
[1,] -0.5238095 0.4285714
[2,] 0.2857143 -0.1428571
2.4. Lists
Lists allow us to combine different data types in a single variable. Lists can be
created using the list() function. This function is similar to the c() function. The
contents of a list are just listed within the list() function as arguments separated by
a comma. List elements can be a vector, matrix or a function. It is possible to name
the elements of the list while creation or later using the names() function.
> L <- list(c(9,1, 4, 7, 0), matrix(c(1,2,3,4,5,6), nrow = 3))
>L
[[1]]
[1] 9 1 4 7 0
[[2]]
[,1] [,2]
[1,] 1 4
[2,] 2 5
[3,] 3 6
> names(L) <- c(“Num”, “Mat”)
>L
$Num
[1] 9 1 4 7 0
$Mat
[,1] [,2]
[1,] 1 4
[2,] 2 5
[3,] 3 6
> L <- list(Num = c(9,1, 4, 7, 0), Mat = matrix(c(1,2,3,4,5,6), nrow = 3))
>L
$Num
[1] 9 1 4 7 0
$Mat
[,1] [,2]
[1,] 1 4
[2,] 2 5
[3,] 3 6
Lists can be nested. That is a list can be an element of another list. But, vectors,
arrays and matrices are not recursive/nested. They are atomic. The functions
is.recursive() and is.atomic() shows if a variable type is recursive or atomic respectively.
> is.atomic(list())
[1] FALSE
45 Data Types in R
> is.recursive(list())
[1] TRUE
> is.atomic(L)
[1] FALSE
> is.recursive(L)
[1] TRUE
> is.atomic(matrix())
[1] TRUE
> is.recursive(matrix())
[1] FALSE
The length() function works on list like in vectors and matrices. But, the dim(),
nrow() and ncol() functions returns only NULL.
> length(L)
[1] 2
> dim(L)
NULL
> nrow(L)
NULL
> ncol(L)
NULL
Arithmetic operations in list are possible only if the elements of the list are of
the same data type. Generally, it is not recommended. As in vectors the elements
of the list can be accessed by indexing them using the square brackets. The index
can be a positive number, or a negative number, or element names or logical values.
> L1 <- list(l1 = c(8, 9, 1), l2 = matrix(c(1,2,3,4), nrow = 2),
l3 = list( l31 = c(“a”, “b”), l32 = c(TRUE, FALSE) ))
> L1
$l1
[1] 8 9 1
$l2
[,1] [,2]
[1,] 1 3
[2,] 2 4
$l3
$l3$l31
[1] “a” “b”
$l3$l32
[1] TRUE FALSE
> L1[1:2]
$l1
[1] 8 9 1
$l2
[,1] [,2]
[1,] 1 3
[2,] 2 4
> L1[-3]
$l1
[1] 8 9 1
$l2
[,1] [,2]
[1,] 1 3
[2,] 2 4
> L1[c(“l1”, “l2”)]

$l1
[1] 8 9 1
47 Data Types in R
$l2
[,1] [,2]
[1,] 1 3
[2,] 2 4
> L1[c(TRUE, TRUE, FALSE)]

$l1
[1] 8 9 1
$l2
[,1] [,2]
[1,] 1 3
[2,] 2 4
A list is a generic vector containing other objects.

> a = c(4,8,12)
> b = c(“abc”, “def ”, “ghi”, “jkl”, “mno”)
> d = c(TRUE, FALSE)
> t = list(a, b, d, 5)
The list t contains copies of the vectors a, b and d. A list slice is retrieved using
single square brackets []. In the below, t[2] contains a slice and a copy of b. Slice can
also be retrieved with multiple members.
> t[2]
[[1]]
[1] “abc” “def ” “ghi” “jkl” “mno”
> t[c(2,4)]
[[1]]
[[2]]
[1] 5
To reference a list member directly double square bracket [[]] is used. Thus
t[[2]] retrieves the second member of the list t. This results in a copy of b, but not a
slice of b. It is also possible to modify the contents of the elements directly, but the
contents of b are unaffected.
> t[[2]]
> t[[2]][1] = “qqq”
> t[[2]]
[1] “qqq” “def ” “ghi” “jkl” “mno”
>b
We can assign names to the list members and reference lists by names instead of
numeric indexes. A list of two members is given as example below with the member
names as “first” and “second”. The list slice containing the member “first” can be
retrieved using the square brackets [] as shown below.
> l = list(first=c(1,2,3), second=c(“a”,”b”, “c”))
>l
$first
[1] 1 2 3
$second
[1] “a” “b” “c”
> l[“first”]
$first
[1] 1 2 3
The named list member can also be directly referenced with the $ operator or
double square brackets [[]] as below.
> l$first
[1] 1 2 3
49 Data Types in R
> l[[“first”]]
[1] 1 2 3
A vector can be converted to a list using the function as.list(). Similarly, a

list can be converted into a vector, provided the list contains scalar elements of
the same type. This is done using the conversion functions such as as.numeric(),
as.character() and so on. If a list consists of non-scalar elements, but if they are of
the same type, then it can be converted into a vector using the function unlist().
> v <- c(7, 3, 9, 2, 6)
> as.list(v)
[[1]]
[1] 7
[[2]]
[1] 3
[[3]]
[1] 9
[[4]]
[1] 2
[[5]]
[1] 6
> L <- list(3, 7, 8, 12, 14)

> as.numeric(L)
[1] 3 7 8 12 14
> L1 <- list(“aaa”, “bbb”, “ccc”)
> L1
[[1]]
[1] “aaa”
[[2]]
[1] “bbb”
[[3]]
[1] “ccc”
> as.character(L1)
[1] “aaa” “bbb” “ccc”
> L1 <- list(l1 = c(78, 90, 21), l2 = c(11,22,33,44,55))
> L1
$l1
[1] 78 90 21
$l2
[1] 11 22 33 44 55
> unlist(L1)
l11 l12 l13 l21 l22 l23 l24 l25
78 90 21 11 22 33 44 55
The c() function can also be used to combine lists as we do for vectors.
> L1 <- list(l1 = c(78, 90, 21), l2 = c(11,22,33,44,55))
> L2 <- list(“aaa”, “bbb”, “ccc”)
> c(L1, L2)
$l1
[1] 78 90 21
$l2
[1] 11 22 33 44 55
[[3]]
[1] “aaa”
[[4]]
[1] “bbb”
[[5]]
[1] “ccc”
51 Data Types in R
2.5. Data Frames

A data frame is used for storing data tables. They store spread-sheet like data. It is a
list of vectors of equal length (not necessarily of the same basic data type). Consider
a data frame df1 consisting of three vectors a, b, and d.
> a = c(1, 2, 3)
> b = c(“a”, “b”, “c”)
> d = c(TRUE, FALSE, TRUE)
> df1 = data.frame(a, b, d)
> df1
a b d
1 1 a TRUE
2 2 b FALSE
3 3 c TRUE
By default the row names are automatically numbered from 1 to the number of
rows in the data frame. It is also possible to provide row names manually using the
row.names argument as below.
> df1 = data.frame(a, b, d, row.names = c(“one”, “two”, “three”))
> df1
a b d
one 1 a TRUE
two 2 b FALSE
three 3 c TRUE
The functions rownames(), colnames(), dimnames(), nrow(), ncol() and dim()

can be applied on the data frames as below. The length() and names() function,
returns the same result as that of ncol() and colnames() respectively.
> rownames(df1)
[1] “one” “two” “three”
> colnames(df1)
[1] “a” “b” “d”

> dimnames(df1)
[[1]]
[1] “one” “two” “three”
[[2]]
[1] “a” “b” “d”
> nrow(df1)
[1] 3
> ncol(df1)
[1] 3
> dim(df1)
[1] 3 3
> length(df1)
[1] 3
> colnames(df1)
[1] “a” “b” “d”
It is possible to create data frames with different length of vectors as long as the
shorter ones can be recycled to match that of the longer ones. Otherwise, an error
will be thrown.
> df2 <- data.frame(x = 1, y = 2:3, y = 4:7)
> df2
x y y.1
1 1 2 4
2 1 3 5
3 1 2 6
4 1 3 7
The argument check.names can be set as FALSE so that a data frame will not
look for valid column names.
53 Data Types in R
> df3 <- data.frame(“BaD col” = c(1:5), “!@#$%^&*” = c(“aaa”))

> df3
BaD.col X........
1 1 aaa
2 2 aaa
3 3 aaa
4 4 aaa
5 5 aaa
There are many built-in data frames available in R (example – mtcars). When
this data frame is invoked in R tool, it produces the below result.
> mtcars
mpg cyl disp hp drat wt ...
Mazda RX4 21.0 6 160 110 3.90 2.62 ...
Mazda RX4 Wag 21.0 6 160 110 3.90 2.88 ...
Datsun 710 22.8 4 108 93 3.85 2.32 ...
............
The top line contains the header or the column names. Each row denotes a record
or a row in the table. A row begins with the name of the row. Each data member of a
row is called a cell. To retrieve a cell value, we enter the row and the column number
of the cell in square brackets [] separated by a comma. The cell value of the second
row and third column is retrieved as below. The row and the column names can also
be used inside the square brackets [] instead of the row and column numbers.
> mtcars[2, 3]
[1] 160
> mtcars[“Mazda RX4 Wag”, “disp”]
[1] 160
The nrow() function gives the number of rows in a data frame and the ncol()
function gives the number of columns in a data frame. To get the preview or the first
few records of a data frame along with the header the head() function can be used.
> nrow(mtcars)
[1] 32
> ncol(mtcars)
[1] 11
> head(mtcars)
mpg cyl disp hp drat wt ...
Mazda RX4 21.0 6 160 110 3.90 2.62 ...
......
To retrieve a column from a data frame we use double square brackets [[]] and
the column name or the column number inside the [[]]. The same can be achieved
by making use of the $ symbol as well. This same result can also be achieved by
using single brackets [] by mentioning a comma instead of the row name / number
and using the column name / number as the second index inside the [].
> mtcars[[“hp”]]
[1] 110 110 93 110 175 105 245 62 95 123 123 180 180 180 ....
> mtcars[[4]]
[1] 110 110 93 110 175 105 245 62 95 123 123 180 180 180 ....
> mtcars$hp
[1] 110 110 93 110 175 105 245 62 95 123 123 180 180 180 ....
> mtcars[,”hp”]
[1] 110 110 93 110 175 105 245 62 95 123 123 180 180 180 ....
> mtcars[,4]
[1] 110 110 93 110 175 105 245 62 95 123 123 180 180 180 ....
Similarly, if we use the column name or the column number inside a single
square bracket [], we get the below result.
> mtcars[4]
hp
Mazda RX4 110
55 Data Types in R
Mazda RX4 Wag 110

Datsun 710 93
....
> mtcars[c(“mpg”,”hp”)]
mpg hp
Mazda RX4 21.0 110
Mazda RX4 Wag 21.0 110
Datsun 710 22.8 93
....
To retrieve a row from a data frame we use the single square brackets [] only by
mentioning the row name / number as the first index inside [] and a comma instead
of the column name / number.
> mtcars[6,]
mpg cyl disp hp drat wt....
Valiant 18.1 6 225 105 2.76 3.46....
> mtcars[c(6,18),]
Valiant 18.1 6 225 105 2.76 3.46....
Fiat 128 32.4 4 78.7 66 4.08 2.20....
> mtcars[“Valiant”,]
Valiant 18.1 6 225 105 2.76 3.46....
> mtcars[c(“Valiant”,”Fiat 128”),]

Valiant 18.1 6 225 105 2.76 3.46....
Fiat 128 32.4 4 78.7 66 4.08 2.20....
If we need to fetch a subset of a data frame by selecting few columns and

specifying conditions on the rows, we can use the subset() function to do this. This
function takes the arguments, the data frame, the condition to be applied on the
rows and the columns to be fetched.
> x <- c(“a”, “b”, “c”, “d”, “e”, “f ”)
> y <- c(3, 4, 7, 8, 12, 15)
> z <- c(TRUE, TRUE, FALSE, TRUE, FALSE, TRUE)
> D <- data.frame(x, y, z)
>D
x y z
1 a 3 TRUE
2 b 4 TRUE
3 c 7 FALSE
4 d 8 TRUE
5 e 12 FALSE
6 f 15 TRUE
> subset(D, y<10 & z, x)

x
1 a
2 b
4 d
As we have for matrices the transpose of a data frame can be obtained using the
t() function as below.
> t(D)
[,1] [,2] [,3] [,4] [,5] [,6]
x “a” “b” “c” “d” “e” “f ”
y “ 3” “ 4” “ 7” “ 8” “12” “15”
z “ TRUE” “ TRUE” “FALSE” “ TRUE” “FALSE” “ TRUE”
57 Data Types in R
The functions rbind() and cbind() can also be applied on the data frames as we
do for the matrices. The only condition for rbind() is that the column names should
match, but for cbind() it does not check even if the column names are duplicated.
> x1 <- c(“aaa”, “bbb”, “ccc”, “ddd”, “eee”, “fff ”)
> y1 <- c(9, 12, 17, 18, 23, 32)
> z1 <- c(TRUE, FALSE, TRUE, FALSE, TRUE, FALSE)
> E <- data.frame(x1, y1, z1)
>E
x1 y1 z1
1 aaa 9 TRUE
2 bbb 12 FALSE
3 ccc 17 TRUE
4 ddd 18 FALSE
5 eee 23 TRUE
6 fff 32 FALSE
> cbind(D, E)
x y z x1 y1 z1
1 a 3 TRUE aaa 9 TRUE
2 b 4 TRUE bbb 12 FALSE
3 c 7 FALSE ccc 17 TRUE
4 d 8 TRUE ddd 18 FALSE
5 e 12 FALSE eee 23 TRUE
6 f 15 TRUE fff 32 FALSE
> F <- data.frame(x, y, z)
>F
x y z
1 a 9 TRUE
2 b 12 FALSE
3 c 17 TRUE
4 d 18 FALSE
5 e 23 TRUE
6 f 32 FALSE
> rbind(D, F)
x y z
1 a 3 TRUE
2 b 4 TRUE
3 c 7 FALSE
4 d 8 TRUE
5 e 12 FALSE
6 f 15 TRUE
7 a 9 TRUE
8 b 12 FALSE
9 c 17 TRUE
10 d 18 FALSE
11 e 23 TRUE
12 f 32 FALSE
The merge() function can be applied to merge two data frames provided they
have common column names. By default, the merge() function does the merging
based on all the common columns, otherwise one of the common column name has
to be specified.
> merge(D, F, by = “x”, all = TRUE)
59 Data Types in R
x y.x z.x y.y z.y

1 a 3 TRUE 9 TRUE
2 b 4 TRUE 12 FALSE
3 c 7 FALSE 17 TRUE
4 d 8 TRUE 18 FALSE
5 e 12 FALSE 23 TRUE
6 f 15 TRUE 32 FALSE
The functions colSums(), colMeans(), rowSums() and rowMeans() can be

applied on the data frames that have numeric values as below.
> x <- c(5, 6, 7, 8)
> y <- c(15, 16, 17, 18)
> z <- c(25, 26, 27, 28)
> G <- data.frame(x, y, z)
>G
x y z
1 5 15 25
2 6 16 26
3 7 17 27
4 8 18 28
> colSums(G[, 1:2])

x y
26 66
> colMeans(G[, 1:3])

x y z
6.5 16.5 26.5
> rowSums(G[1:3, ])
1 2 3
45 48 51
> rowMeans(G[2:4, ])
2 3 4
16 17 18
2.6. Factors
Factors are used to store categorical data like gender (“Male” or “Female”). They
behave sometimes like character vectors and sometimes like integer vectors based
on the context.
Factors stores categorical data and they behave like strings sometimes and
integers sometimes. Consider a data frame that stores the weight of few males
and females. In this case the column that stores the gender is a factor as it stores
categorical data. The choices “female” and “male” are called the levels of the factor.
This can be viewed by using the levels() function and nlevels() function.
> weight <- data.frame(wt_kg = c(60,82,45, 49,52,75,68),
gender = c(“female”,”male”, “female”, “female”, “female”, “male”, “male”))
> weight
wt_kg gender
1 60 female
2 82 male
3 45 female
4 49 female
5 52 female
6 75 male
7 68 male
> weight$gender
61 Data Types in R
[1] female male female female female male male

Levels: female male
> levels(weight$gender)
[1] “female” “male”
> nlevels(weight$gender)
[1] 2
At the atomic level a factor can be created using the factor() function, which
takes a character vector as the argument.
> gender <- factor(c(“female”, “male”, “female”, “female”, “female”, “male”, “male”))
> gender
Levels: female male
The levels argument can be used in the factor() function to specify the levels of
the factor. It is also possible to change the levels once the factor is created. This is
done using the function levels() or the function relevel(). The function relevel() just
mentions which level comes first.
> gender <- factor(c(“female”, “male”, “female”, “female”, “female”,
“male”, “male”), levels = c(“male”, “female”))
> gender
Levels: male female
> levels(gender) <- c(“F”, “M”)
> gender
[1] M F M M M F F
Levels: F M
> relevel(gender, “M”)
[1] M F M M M F F
Levels: M F
It is possible to drop a level from a factor using the function droplevels() when
the level is not in use as in the example below. [Note: the function is.na() is used to
remove the missing value].
> diet <- data.frame(eat = c(“fruit”, “fruit”, “vegetable”, “fruit”),
type = c(“apple”, “mango”, NA, “papaya”))
> diet
eat type
1 fruit apple
2 fruit mango
3 vegetable <NA>
4 fruit papaya
> diet <- subset(diet, !is.na(type))
> diet
eat type
1 fruit apple
2 fruit mango
4 fruit papaya
> diet$eat
[1] fruit fruit fruit
Levels: fruit vegetable
> levels(diet)
NULL
> levels(diet$eat)
[1] “fruit” “vegetable”
> unique(diet$eat)
[1] fruit
Levels: fruit vegetable
63 Data Types in R
> diet$eat <- droplevels(diet$eat)

> levels(diet$eat)
[1] “fruit”
In some cases, the levels need to be ordered as in rating a product or course. The
ratings can be “Outstanding”, “Excellent”, “Very Good”, “Good”, “Bad”. When a
factor is created with these levels, it is not necessary they are ordered. So, to order the
levels in a factor, we can either use the function ordered() or the argument ordered =
TRUE in the factor() function. Such ordering can be useful when analysing survey
data.
> ch <- c(“Outstanding”, “Excellent”, “Very Good”, “Good”, “Bad”)
> val <- sample(ch, 100, replace = TRUE)
> rating <- factor(val, ch)
> rating
[1] Outstanding Bad Outstanding Good Very Good Very Good
[7] Excellent Outstanding Bad Excellent Very Good Bad
...
Levels: Outstanding Excellent Very Good Good Bad
> is.factor(rating)
[1] TRUE
> is.ordered(rating)
[1] FALSE
> rating_ord <- ordered(val, ch)
> is.factor(rating_ord)
[1] TRUE
> is.ordered(rating_ord)
[1] TRUE
> rating_ord
[1] Outstanding Bad Outstanding Good Very Good Very Good

[7] Excellent Outstanding Bad Excellent Very Good Bad
...
Levels: Outstanding < Excellent < Very Good < Good < Bad
Numeric values can be summarized into factors using the cut() function and
the result can be viewed using the table() function which lists the count of numbers
in each category. For example let us consider the variable age which has the numeric
values of ages. These ages can be grouped using the cut() function with an interval
of 10 and the result is a factor age_group.
> age <- c(18,20, 31, 32, 33, 35, 41, 38, 45, 48, 51, 27, 29, 42, 39)
> age_group <- cut(age, seq.int(15, 55, 10))
> age
[1] 18 20 31 32 33 35 41 38 45 48 51 27 29 42 39
> age_group
[1] (15,25] (15,25] (25,35] (25,35] (25,35] (25,35] (35,45] (35,45] (35,45] (45,55]
[11] (45,55] (25,35] (25,35] (35,45] (35,45]
Levels: (15,25] (25,35] (35,45] (45,55]
> table(age_group)
age_group
(15,25] (25,35] (35,45] (45,55]
2 6 5 2
The function gl() can be used to create a factor, which takes the first argument
that tells how many levels the factor contains and the second argument that tells
how many times each level has to be repeated as value. This function can also take
the argument labels, which lists the names of the factor levels. The function can
also be made to list alternating values of the labels as below.
> gl(5,3)
[1] 1 1 1 2 2 2 3 3 3 4 4 4 5 5 5
Levels: 1 2 3 4 5
65 Data Types in R
> gl(5,3, labels = c(“one”, “two”, “three”, “four”, “five”))

[1] one one one two two two three three three four four four five
[14] five five
Levels: one two three four five
> gl(5,1,15)
[1] 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5
Levels: 1 2 3 4 5
The factors thus generated can be combined using the function interaction() to
get a resultant combined factor.
> fac1 <- gl(5,3, labels = c(“one”, “two”, “three”, “four”, “five”))
> fac2 <- gl(5,1,15, labels = c(“a”, “b”, “c”, “d”, “e”, “f ”, “g”, “h”, “i”, “j”,
“k”, “l”, “m”, “n”, “o”))
> interaction(fac1, fac2)
[1] one.a one.b one.c two.d two.e two.a three.b three.c three.d four.e
[11] four.a four.b five.c five.d five.e
75 Levels: one.a two.a three.a four.a five.a one.b two.b three.b four.b ... five.o
2.7. Strings
Strings are stored in character vectors. Most string manipulation functions act
on character vectors. Character vectors can be created using the c() function by
enclosing the string in double or single quotes. (Generally we follow only double
quotes). The paste() function can be used to concatenate two strings with a space
in between. If the space need not be shown, we use the function paste0(). To have
specified separator between the two concatenated string, we use the argument
sep in the paste() function. The result can be collapsed into one string using the
collapse argument.
> c(“String 1”, ‘String 2’)
[1] “String 1” “String 2”
> paste(c(“Pine”, “Red”), “Apple”)
[1] “Pine Apple” “Red Apple”

> paste0(c(“Pine”, “Red”), “Apple”)
[1] “PineApple” “RedApple”
> paste(c(“Pine”, “Red”), “Apple”, sep = “-”)
[1] “Pine-Apple” “Red-Apple”
> paste(c(“Pine”, “Red”), “Apple”, sep = “-”, collapse = “, “)
[1] “Pine-Apple, Red-Apple”
The to String() function can be used to convert a number vector into a character
vector, with the elements separated by a comma and a space. It is possible to specify
the width of the print string in this function.
> x <- c(1:10)^3
>x
[1] 1 8 27 64 125 216 343 512 729 1000
> toString(x)
[1] “1, 8, 27, 64, 125, 216, 343, 512, 729, 1000”
> toString(x, 18)
[1] “1, 8, 27, 64, ....”
The cat() function is also similar to the paste() function, but there is little
difference in it as shown below.
> cat(c(“Red”, “Pine”), “Apple”)
Red Pine Apple
The noquote() function forces the string outputs not to be displayed with
quotes.
> a <- c(“I”, “am”, “a”, “data”, “scientist”)
>a
[1] “I” “am” “a” “data” “scientist”
> noquote(a)
[1] I am a data scientist
67 Data Types in R
The formatC() function is used to format the numbers and display them as
strings. This function has the arguments digits, width, format, flag etc which can be
used as below. A slight variation of the function formatC() is the function format()
whose usage is as shown below.
> h <- c(4.567, 8.981, 27.772)
>h
[1] 4.567 8.981 27.772
> formatC(h)
[1] “4.567” “8.981” “27.77”
> formatC(h, digits = 3)
[1] “4.57” “8.98” “27.8”
> formatC(h, digits = 3, width = 5)
[1] “ 4.57” “ 8.98” “ 27.8”
> formatC(h, digits = 3, format = “e”)
[1] “4.567e+00” “8.981e+00” “2.777e+01”
> formatC(h, digits = 3, flag = “+”)
[1] “+4.57” “+8.98” “+27.8”
> format(h)
[1] “ 4.567” “ 8.981” “27.772”
> format(h, digits = 3)
[1] “ 4.57” “ 8.98” “27.77”
> format(h, digits = 3, trim = TRUE)
[1] “4.57” “8.98” “27.77”
The sprint() function is also used for formatting strings and passing number
values in between the strings. The argument %s in this function stands for a string
to be passed. The argument %d and argument %f stands for integer and floating-
point number. The usage of this function can be understood by the below example.
> x <- c(1, 2, 3)

> sprintf(“The number %d in the list is = %f ”, x, h)
[1] “The number 1 in the list is = 4.567000”
To print a tab in between text, we can use the cat() function with the special
character “\t” included in between the text as below. Similarly, if we need to insert
a new line in between the text, we use “\n”. In this cat() function the argument fill
= TRUE means that after printing the text, the cursor is placed in the next line.
Suppose if a back slash has to be used in between the text, it is preceded by another
back slash. If we enclose the text in double quotes and if the text contains a double
quote in between, it is also preceded by a back slash. Similarly, if we enclose the
text in single quotes and if the text contains a single quote in between, it is also
preceded by a back slash. If we enclose the text in double quotes and if the text
contains a single quote in between, or if we enclose the text in single quotes and if
the text contains a double quote in between, it is not a problem (No need for back
slash).
> cat(“Black\tBerry”, fill = TRUE)
Black Berry
> cat(“Black\nBerry”, fill = TRUE)
Black
Berry
> cat(“Black\\Berry”, fill = TRUE)
Black\Berry
> cat(“Black\”Berry”, fill = TRUE)
Black”Berry
> cat(‘Black\’Berry’, fill = TRUE)
Black’Berry
> cat(‘Black”Berry’, fill = TRUE)
Black”Berry
69 Data Types in R
> cat(“Black’Berry”, fill = TRUE)

Black’Berry
The function toupper() and tolower() are used to convert a string into upper
case or lower case respectively. The substring() or the substr() function is used to cut
a part of the string from the given text. Its arguments are the text, starting position
and ending position. Both these functions produce the same result.
> toupper(“The cat is on the Wall”)
[1] “THE CAT IS ON THE WALL”
> tolower(“The cat is on the Wall”)
[1] “the cat is on the wall”
> substring(“The cat is on the wall”, 3, 10)

[1] “e cat is”
> substr(“The cat is on the wall”, 3, 10)
[1] “e cat is”
> substr(“The cat is on the wall”, 5, 10)
[1] “cat is”
The function strsplit() does the splitting of a text into many strings based on
the splitting character mentioned as argument. In the below example the splitting
is done when a space is encountered. It is important to note that this function
returns a list and not a character vector as a result.
> strsplit(“I like Bannana, Orange and Pineapple”, “ “)
[[1]]
[1] “I” “like” “Bannana,” “Orange” “and” “Pineapple”
In this same example if the text has to be split when a comma or space is
encountered it is mentioned as “,?”. This means that the comma is optional and
space is mandatory for splitting the given text.
> strsplit(“I like Bannana, Orange and Pineapple”, “,? “)

[[1]]
[1] “I” “like” “Bannana” “Orange” “and” “Pineapple”
The default R’s working directory can be obtained using the function getwd()
and this default directory can be changed using the function setwd(). The directory
path mentioned in the setwd() function should have the forward slash instead of
backward slash as in the example below.
> getwd()
> setwd(“C:/Program Files/R”)
> getwd()
[1] “C:/Program Files/R”
It is also possible to construct the file paths using the file.path() function which
automatically inserts the forward slash between the directory names. The function
R.home() list the home directory where R is installed.
> file.path(“C:”, “Program Files”, “R”, “R-3.3.0”)
[1] “C:/Program Files/R/R-3.3.0”
> R.home()
[1] “C:/PROGRA~1/R/R-33~1.0”
Paths can also be specified by relative terms such as “.” denotes current directory,
“..” denotes parent directory and “~” denotes home directory. The function path.
expand() converts relative paths to absolute paths.
> path.expand(“.”)
[1] “.”
> path.expand(“..”)
[1] “..”
> path.expand(“~”)
71 Data Types in R
The function basename() returns only the file name leaving its directory if
specified. On the other hand the function dirname() returns only the directory
name leaving the file name.
> filename <- “C:/Program Files/R/R-3.3.0/bin/R.exe”
> basename(filename)
[1] “R.exe”
> dirname(filename)
[1] “C:/Program Files/R/R-3.3.0/bin”
2.8. Dates and Times

Dates and Times are common in data analysis and R has a wide range of capabilities
for dealing with dates and times.
2.8.1. Date and Time Classes
R has three date and time base classes and they are POSIXct, POSIXlt and Date.
POSIX is a set of standards that defines how dates and times should be specified and
“ct” stands for “calendar time”. POSIXlt stores dates as a list of seconds, minutes,
hours, day of month etc. For storing and calculating with dates, we can use POSIXct
and for extracting parts of dates, we can use POSXlt.
The function Sys.time() is used to return the current date and time. This
returned value is by default in the POSIXct form. But, this can be converted to
POSIXlt form using the function as.POSIXlt(). When printed both forms of date
and time are displayed in the same manner, but their internal storage mechanism
differs. We can also access individual components of a POSIXlt date using the dollar
symbol or the double brackets as shown below.
> Sys.time()
[1] “2017-05-11 14:31:29 IST”
> t <- Sys.time()
> t1 <- Sys.time()
> t2 <- as.POSIXlt(t1)
> t1
[1] “2017-05-11 14:39:39 IST”
> t2
[1] “2017-05-11 14:39:39 IST”
> class(t1)
[1] “POSIXct” “POSIXt”
> class(t2)
[1] “POSIXlt” “POSIXt”
> t2$sec
[1] 39.20794
> t2[[“min”]]
[1] 39
> t2$hour
[1] 14
> t2$mday
[1] 11
> t2$wday
[1] 4
The Date class stores the dates as number of days from start of 1970. This class
is useful when time is insignificant. The as.Date() function can be used to convert
a date in other class formats to the Date class format.
> t3 <- as.Date(t2)
> t3
[1] “2017-05-11”
There are also other add-on packages available in R to handle date and time and
they are date, dates, chron, yearmon, yearqtr, timeDate, ti and jul.
73 Data Types in R
2.8.2. Date Conversions
In CSV files the dates will be normally stored as strings and they have to be converted
into date and time using any of the packages. For this we need to parse the strings
using the function strptime() and this returns the date of the format POSIXlt.
The date format is specified as a string and passed as argument to the strptime()
function. If the given string does not match the format given in the format string,
then it returns NA.
> date1 <- strptime(“22:15:45 22/08/2015”, “%H:%M:%S %d/%m/%Y”)
> date1
[1] “2015-08-22 22:15:45 IST”
> date2 <- strptime(“22:15:45 22/08/2015”, “%H:%M:%S %d-%m-%Y”)

> date2
[1] NA
In the format string “%H” denotes hour in 24 hour system, “%M” denotes
minutes, “%S” denotes second, “%m” denotes the number of the month, “%d”
denotes the day of the month as number, “%Y” denotes four digit year.
To convert a date into a string the function strftime() is used. This function also
takes a date formatting string as argument like strptime(). In the format string “%I”
denotes hour in 12 hours system, “%p” denotes AM/PM, “%A” denotes the string of
day of the week, “%B” denotes the string of name of the month.
> strftime(Sys.Date(),”It’s %I:%M%p on %A %d %B, %Y.”)
[1] “It’s 12:00AM on Thursday 11 May, 2017.”
2.8.3. Time Zones
It is possible to specify the time zone when parsing a date string using strptime()
or strftime() functions. If this is not specified, the default time zone is taken. The
functions Sys.timezone() and Sys.getlocale(“LC_TIME”) are used to get the default
time zone of the system and the operating system respectively.
> Sys.timezone()
[1] “Asia/Calcutta”
> Sys.getlocale(“LC_TIME”)
[1] “English_India.1252”
Few of the time zones are UTC (Universal Time), IST (Indian Standard Time),
EST (Eastern Standard Time), PST (Pacific Standard Time), GMT (Greenwitch
Meridian Time), etc. It is also possible to give manual offset from UTC as “UTC+n”
or “UTC–n” to denote west and east parts of UTC respectively. Even though it
throws warning message, it gives the result correctly.
> strftime(Sys.time(), tz = “UTC”)
[1] “2017-05-12 04:59:04”
> strftime(Sys.time(), tz = “UTC-5”)

[1] “2017-05-12 09:59:09”
Warning message:
In as.POSIXlt.POSIXct(x, tz = tz) : unknown timezone ‘UTC-5’
> strftime(Sys.time(), tz = “UTC+5”)
[1] “2017-05-11 23:59:15”
Warning message:
In as.POSIXlt.POSIXct(x, tz = tz) : unknown timezone ‘UTC+5’
The time zone changes does not happen in strftime() function if the date is in
POSIXlt dates. Hence, it is required to change to POSIXct format first and then
apply the function.
2.8.4. Calculations with Dates and Times
If we add a number to the POSIXct or POSIXlt classes, it will shift to that many
seconds. If we add a number to the Date class, it will shift to that many days.
> ct <- as.POSIXct(Sys.time())
> lt <- as.POSIXlt(Sys.time())
75 Data Types in R
> dt <- as.Date(Sys.time())

> ct
[1] “2017-05-12 11:41:54 IST”
> ct + 2500
[1] “2017-05-12 12:23:34 IST”
> lt
[1] “2017-05-12 11:42:15 IST”
> lt + 2500
[1] “2017-05-12 12:23:55 IST”
> dt
[1] “2017-05-12”
> dt + 2
[1] “2017-05-14”
Adding two dates, throws error. But subtracting two dates gives the number
of days in between the dates. To get the same result, alternatively, the difftime()
function can be used and in this it is possible to specify the attribute units = “secs”
(or “mins” or “hours” or “days” or “weeks”).
> dt1 <- as.Date(“10/10/1973”, “%d/%m/%Y”)
> dt1
[1] “1973-10-10”
> dt2 <- as.Date(“25/09/2000”, “%d/%m/%Y”)
> dt2
[1] “2000-09-25”
> diff <- dt2 - dt1

> diff
Time difference of 9847 days
> difftime(dt2, dt1)


> difftime(dt2, dt1, units = “secs”)
Time difference of 850780800 secs
> difftime(dt2, dt1, units = “mins”)
Time difference of 14179680 mins
> difftime(dt2, dt1, units = “hours”)
Time difference of 236328 hours
> difftime(dt2, dt1, units = “days”)
> difftime(dt2, dt1, units = “weeks”)
Time difference of 1406.714 weeks
The seq() function can be used to generate a sequence of dates. The argument
“by” can take many options based on the class of the dates specified. We can apply
the mean() and summary() functions on these sequence of dates generate.
> seq(dt1, dt2, by = “1 year”)
[1] “1973-10-10” “1974-10-10” “1975-10-10” “1976-10-10” “1977-10-10”
“1978-10-10”
[7] “1979-10-10” “1980-10-10” “1981-10-10” “1982-10-10” “1983-10-10”
“1984-10-10”
[13] “1985-10-10” “1986-10-10” “1987-10-10” “1988-10-10” “1989-10-10”
“1990-10-10”
[19] “1991-10-10” “1992-10-10” “1993-10-10” “1994-10-10” “1995-10-10”
“1996-10-10”
[25] “1997-10-10” “1998-10-10” “1999-10-10”
> seq(dt1, dt2, by = “500 days”)

[1] “1973-10-10” “1975-02-22” “1976-07-06” “1977-11-18” “1979-04-02”
“1980-08-14”
77 Data Types in R
[7] “1981-12-27” “1983-05-11” “1984-09-22” “1986-02-04” “1987-06-19”

“1988-10-31”
[13] “1990-03-15” “1991-07-28” “1992-12-09” “1994-04-23” “1995-09-05”
“1997-01-17”
[19] “1998-06-01” “1999-10-14”
> mean(seq(dt1, dt2, by = “1 year”))

[1] “1986-10-10”
> summary(seq(dt1, dt2, by = “1 year”))

Min. 1st Qu. Median Mean 3rd Qu. Max.
“1973-10-10” “1980-04-10” “1986-10-10” “1986-10-10” “1993-04-10” “1999-10-10”
The lubridate package makes the process of date and time manipulation easier.
The ymd() function in this package converts any date to the format of year, month
and day separated by hyphens.(Note: This function requires the date to be specified
in the order of year, month and day, but can use any separator as below).
> install.packages(“lubridate”)
> library(lubridate)
> ymd(“2000/09/25”, “2000-9-25”, “2000*9.25”)
[1] “2000-09-25” “2000-09-25” “2000-09-25”
If the given date is in other formats that is not in the order of year, month and
day, then we have other functions such as ydm(), mdy(), myd(), dmy() and dym().
These functions can also be accompanied with time by making use of the functions
ymd_h(), ymd_hm() and ymd_hms() [similar functions available for ydm(), mdy(),
myd(), dmy() and dym()]. All the parsing functions in the lubridate package returns
POSIXct dates and the default time zone is UTC. A function named stamp() in the
lubridate package allows formatting of the dates in a human readable format.
> dt_format <- stamp(“I purchased on Sunday, the 10th of October 2013 at
6:00:00 PM”)
Multiple formats matched: “I purchased on %A, the %dth of %B %Y at %H:%M:%S
%Op”(0), “I purchased on %A, the %dth of October %Y at %Om:%H:%M %Op”...

...
Using: “I purchased groceries on %A, the %dth of %Om %Y at %H:%M:%S %Op”
> dt_to_convert <- strptime(“2000-09-25 7:00:00”, “%Y-%m-%d %H:%M:%S”)

> dt_format(dt_to_convert)
[1] “I purchased groceries on Monday, the 25th of 09 2000 at 07:00:00 AM”
The lubridate package has three variable types, namely the “Durations”, “Periods”
and “Intervals”. The lubridate package has the functions, dyears(), dweeks(), ddays(),
dhours(), dminutes(), dseconds() etc that specify the duration of year, week, day, hour,
minute and second in terms of seconds. The duration of 1 minute is 60 seconds,
the duration of 1 hour is 3600 seconds (60 minutes * 60 seconds), the duration of 1
day is 86,400 seconds (24 hours * 60 minutes * 60 seconds), the duration of 1 year
is 31,536,000 seconds (365 days * 24 hours * 60 minutes * 60 seconds) and so on.
The function today() returns the current days date.
> y <- dyears(1:5)
>y
[1] “31536000s (~52.14 weeks)” “63072000s (~2 years)” “94608000s (~3 years)”
[4] “126144000s (~4 years)” “157680000s (~5 years)”
> w <- dweeks(1:4)
>w
[1] “604800s (~1 weeks)” “1209600s (~2 weeks)” “1814400s (~3 weeks)”
[4] “2419200s (~4 weeks)”
> d <- ddays(1:10)
>d
[1] “86400s (~1 days)” “172800s (~2 days)” “259200s (~3 days)”
[4] “345600s (~4 days)” “432000s (~5 days)” “518400s (~6 days)”
[7] “604800s (~1 weeks)” “691200s (~1.14 weeks)” “777600s (~1.29 weeks)”
[10] “864000s (~1.43 weeks)”
79 Data Types in R
> today() + y
[1] “2018-05-12” “2019-05-12” “2020-05-11” “2021-05-11” “2022-05-11”
“Periods” specify time spans according to the clock time. The lubridate package
has the functions, years(), weeks(), days(), hours(), minutes(), seconds() etc that
specify the period of year, week, day, hour, minute and second in terms of clock
time. The exact length of these periods can be realized only if they are added to an
instance of date or time.
> y <- years(1:7)
>y
[1] “1y 0m 0d 0H 0M 0S” “2y 0m 0d 0H 0M 0S” “3y 0m 0d 0H 0M 0S”
“4y 0m 0d 0H 0M 0S”
[5] “5y 0m 0d 0H 0M 0S” “6y 0m 0d 0H 0M 0S” “7y 0m 0d 0H 0M 0S”
> today()+y
[1] “2018-05-12” “2019-05-12” “2020-05-12” “2021-05-12” “2022-05-12”
“2023-05-12”
[7] “2024-05-12”
“Intervals” are defined by the instance of date or time at the beginning and
end. They are mostly used for specifying “Periods” and “Durations” and conversion
between “Periods” and “Durations”.
> yr <- dyears(5)
> yr
[1] “157680000s (~5 years)”
> as.period(yr)
[1] “5y 0m 0d 0H 0M 0S”
> sdt <- ymd(“2017-05-12”)
> int <- new_interval(sdt, sdt+yr)
> int
[1] 2017-05-12 UTC--2022-05-11 UTC
The operator “%--%” is used for defining intervals and the operator “%within%”
is used for checking if a given date is within the given interval.
> intv <- ymd(“1973-10-10”) %--% ymd(“2000-09-25”)
> intv
[1] 1973-10-10 UTC--2000-09-25 UTC
> ymd(“1979-12-12”) %within% intv
[1] TRUE
The function with_tz() can be used to change the time zone of a date (correctly
handles POSIXlt dates) and the function force_tz() is used for updating incorrect
time zones.
> with_tz(Sys.time(), tz = “America/Los_Angeles”)
[1] “2017-05-12 06:44:14 PDT”
> with_tz(Sys.time(), tz = “Asia/Kolkata”)
[1] “2017-05-12 19:14:29 IST”
The functions floor_date() and ceiling_date() can be used to find the lower and
upper limit of a given date as below.
> floor_date(today(), “year”)
[1] “2017-01-01”
> ceiling_date(today(), “year”)
[1] “2018-01-01”
> floor_date(today(), “month”)
[1] “2017-05-01”
> ceiling_date(today(), “month”)
[1] “2017-06-01”
 HIGHLIGHTS
 The basic data types in R are Numeric, Integer, Complex, Logical and
Character.
81 Data Types in R
 To create an integer variable and force a numeric value into an integer in

R, the as.integer() function is invoked.
 A paste() function can be used to concatenate two character values.
 The substr() function can be used to extract a substring from a given
string.
 The sub() function is used to replace the first occurrence of a string with
another string.
 The vector() function creates a vector of a specified type and length.
 The seq() function allows to generate sequences.
 The function length() can be used to find the length of the vector.
 The which() function returns the elements of the vector which satisfies
the condition specified within this function.
 To create an array the array() function can be used.
 A matrix can also be created using the array() function where the
dimension of the array is two.
 The columns of two matrices can be combined using the cbind() function.
 The rows of two matrices can be combined using the rbind() function.
 Lists can be created using the list() function.
 To get the preview or the first few records of a data frame along with the
header the head() function can be used.
 The merge() function can be applied to merge two data frames provided
they have common column names.
 Factors are used to store categorical data and they behave like strings
sometimes and integers sometimes.
 At the atomic level a factor can be created using the factor() function.
 The function gl() can be used to create a factor.
 The factors can be combined using the function interaction().
 The default R’s working directory can be obtained using the function
getwd().
 The default directory can be changed using the function setwd().
 The function R.home() list the home directory where R is installed.
 R has three date and time base classes and they are POSIXct, POSIXlt and
Date.
 The function Sys.time() is used to return the current date and time.
 The as.Date() function can be used to convert a date in other class formats
to the Date class format.
 The functions Sys.timezone() is used to get the default time zone of the
system.
 The lubridate package has the functions, dyears(), dweeks(), ddays(),
dhours(), dminutes(), dseconds() etc that specify the duration of year,
week, day, hour, minute and second in terms of seconds.
 The lubridate package has the functions, years(), weeks(), days(), hours(),
minutes(), seconds() etc that specify the period of year, week, day, hour,
minute and second in terms of clock time.
CHAPTER 3
Data Preparation
 OBJECTIVES

 know about the default datasets available in R
 know how to import and export CSV files in R
 know how to import unstructured data files into R
 know how to import XML and HTML files into R
 know how to import JASON and YAML files into R
 know how to import and export excel files in R
 know how to import SAS, SPSS and MATLAB files into R
 know how to import web data files into R
 understand the concept of accessing various databases from R
 manipulate string data
 manipulate data frames
 understand how to melt and cast data in data frames
 understand how the grouping functions are applied on the data in R
3.1. Datasets
R has many datasets built in. R can read data from variety of other data sources
and in variety of formats. One of the packages in R is datasets which is filled with
example datasets. Many other packages also contain datasets. We can see all the
datasets available in the loaded packages using the data() function.
To access a particular dataset use the data() function with its argument as the
dataset name enclosed within double quotes and the second optional argument
being the package name in which the dataset is present (This second argument is
required only if the particular package is not loaded). The invoked dataset can be
listed just like a data frame using the head() function.
> data(“kidney”, package = “survival”)
> head(kidney)
id time status age sex disease frail
1 1 8 1 28 1 Other 2.3
2 1 16 1 28 1 Other 2.3
3 2 23 1 48 2 GN 1.9
….
Figure 3.1 R-Studio Showing the List of Datasets

85 Data Preparation
3.2. Importing and Exporting Files
3.2.1. Text and CSV Files
Text documents have several formats. Common format are CSV (Comma Separated
Values), XML (Extended Markup Language), JSON (JavaScript Object Notation)
and YAML. An example of an unstructured text data is a book.
Comma Separated Values (CSV) Files is a spreadsheet like data stored with
comma delimited values. The read.table() function reads these files and stores the
result in a data frame. If the data has header, it is required to pass the argument
header = TRUE to the read.table() function. The argument fill = TRUE makes
the read.table() function substitute NA values for the missing fields. The system.
file() function is used to locate files that are inside a package. In the below example
“extdata” is the folder name and the package name is “learning” and the file name
is “RedDeerEndocranialVolume,dlm” The str() function takes the data frame name
as the argument and lists the structure of the dataset stored in the data frame.
> install.packages(“learningr”)
> library(learningr)
> deer_file <- system.file(“extdata”,”RedDeerEndocranialVolume.dlm”,
package = “learningr”)
> deer_data <- read.table(deer_file, header=TRUE, fill=TRUE)
> str(deer_data)
‘data.frame’: 33 obs. of 8 variables:
$ SkullID : Factor w/ 33 levels “A4”,”B11”,”B12”,..: 14 2 17 16 15 13 10 11
19 3 ...
$ VolCT : int 389 389 352 388 375 325 346 302 379 410 ...
$ VolBead : int 375 370 345 370 355 320 335 295 360 400 ...
$ VolLWH : int 1484 1722 1495 1683 1458 1363 1250 1011 1621 1740 ...
$ VolFinarelli: int 337 377 328 377 328 291 289 250 347 387 ...
$ VolCT2 : int NA NA NA NA NA NA 346 303 375 413 ...
$ VolBead2 : int NA NA NA NA NA NA 330 295 365 395 ...

$ VolLWH2 : int NA NA NA NA NA NA 1264 1009 1647 1728 ...
The column names and row names are listed by default and if the row names
are not given in the dataset, the rows are simply numbered 1, 2, 3 and so on. The
arguments specify how the file will be read. The argument sep determines the
character to use as separator between fields. The nrow argument specifies the lines
of data to read. The argument skip specifies the number of lines to skip at the start
of the file. For the functions read.table() and read.csv() the default separator is set
to comma and they assume the data has header row. The function read.csv2() uses
the semicolon as the separator and comma instead of decimals. The read.delim()
function imports the tab-delimited files with full stops for decimal places. The read.
delim2() function imports the tab-delimited files with commas for decimal places.
> read.csv(deer_file, header=FALSE, skip = 3, nrow = 2)
V1
1 DIC90 352 345 1495 328
2 DIC83 388 370 1683 377
> head(deer_data)
SkullID VolCT VolBead VolLWH VolFinarelli VolCT2 VolBead2 VolLWH2
1 DIC44 389 375 1484 337 NA NA NA
2 B11 389 370 1722 377 NA NA NA
3 DIC90 352 345 1495 328 NA NA NA
….
The colbycol and sqldf packages contain functions that allow to read part of
the CSV file into R. These are useful when we don’t need all the columns or all the
rows. For low-level control we can use the scan() function to import CSV file. For
data exported from other languages we may need to pass the na.strings argument to
the read.table() function to replace the missing values. If the data is exported from
SQL, we use na.strings = “NULL” and if the data is exported from SAS or Stata,
we use na.strings = “.”. If the data is exported from Excel we use the na.strings =
c(“”,”#N/A”, “#DIV/0!”, “#NUM!”).
87 Data Preparation
Writing data from R into a file is easier than reading files into R. For this we use
the functions write.table() and write.csv(). These functions take a data frame and a
file path as arguments. They also have arguments to specify if we need not include
row names in the output file or to specify the character encoding of the output file.
> write.csv(deer_data,”F:/deer.csv”, row.names = FALSE, fileEncoding = “utf8”)
3.2.2. Unstructured Files
If the file structure is week, it is easier to read the file as lines of text using the function
readLines() and then parse the contents. The readLines() function accepts a path
to the file as the argument. Similarly, the writeLines() function takes a text line or a
character vector and the file name as argument and writes the text to the file.
> tempest <- readLines(“F:/Tempest.txt”)
> tempest
[1] “The writing of Prefaces to Plays was probably invented by some very”
[2] “ambitious Poet, who never thought he had done enough: Perhaps by
ome”
[3] “Ape of the French Eloquence, which uses to make a business of a Letter
of ”
....
> writeLines(“This book is about a story by Shakespeare”, “F:/story.csv”)
3.2.3. XML and HTML Files
XML files are used for storing nested data. Few of them are RSS (Really Simple
Syndication) feeds, SOAP (Simple Object Access Protocols) and XHTML Web Pages.
To read the XML files, the XML package has to be installed. When an XML file is
imported, the result can be stored using the internal nodes or the R nodes. If the result
is stored using internal nodes, it allows to query the node tree using the XPath language
(used for interrogating XML documents). The XML file can be imported using the
function xmlParse() function. This function can take the argument useInternalNodes
= FALSE to use R-level nodes instead of the internal nodes while importing the XML
files. But, this is set by default by the xml TreeParse() function.
> install.packages(“XML”)
> library(XML)
> xml_file <- system.file(“extdata”, “options.xml”, package = “learningr”)

> r_options <- xmlParse(xml_file)
> xmlParse(xml_file, useInternalNodes = FALSE)

> xmlTreeParse(xml_file)
The functions for importing HTML pages are htmlParse() and htmlTreeParse()
and they behave same as the xmlParse() and xmlTreeParse() functions.
3.2.4. JASON and YAML Files
The two packages dealing with JSON data are RJSONIO and rjson and the best of
these is the RJSONIO. The function used to import the JSON file is fromJSON()
and the function used to export the JSON file is toJSON(). The yaml package has
two functions for importing YAML data and they are yaml.load() and yaml.load_
file(). The function as.yaml() performs the task of converting R objects to YAML
strings.
Many softwares store their data in binary formats which are smaller in size
than the text files. They hence provide performance gains at the expense of human
readability.
3.2.5. Excel Files
Excel is the world’s most powerful data analysis tool and its document formats
are XLX and XLSX. Spreadsheets can be imported with the functions read.xlsx()
and read.xlsx2(). The colClasses argument determines what class each column
should have in the resulting data frame and this argument is optional in the above
functions. To write to an excel file from R we use the function write.xlsx2() that
takes the data frame and the file name as arguments. There is another package
xlsReadWrite that does the same function of the xlxs package but this one works
only in 32-bit R installations and only on windows.
89 Data Preparation
> install.packages(“xlsx”)
> library(xlsx)
> logfile <- read.xlsx2(“F:/Log2015.xls”, sheetIndex = 1, startRow = 2, endrow = 72,
colIndex = 1:5, colClasses = c(“character”, “numeric”, “character”,
“character”, “integer”))
3.2.6. SAS, SPSS and MATLAB Files
The files from a statistical package are imported using the foreign package. The
read.ssd() function is used to read SAS datasets and the read.dta() function is
used to read Stata DTA files. The read.spss() function is used to import the SPSS
data files. Similarly, these files can be written with the write.foreign() function.
The MATLAB binary data files can be read and written using the readMat() and
writeMat() functions in the R.matlab package. The files in picture formats can be
read via the jpeg, png, tiff, rtiff and readbitmap packages.
3.2.7. Web Data
R has ways to import data from web sources using Application Programming Interface
(API). For example the World Bank makes its data available using the WDI package
and the Polish government data can be accessed using the SmarterPoland package.
The twitter package provides access to Twitter’s users and their tweet.
The read.table() function can accept URL rather than a local file. Accessing a
large file from internet can be slow and if the file is required frequently, it is better
to download the file using the download.file() function and create a local copy and
then import that.
> cancer_url <- “http://repository.seasr.org/Datasets/UCI/csv/breast-cancer.csv”
> cancer_data <- read.csv(cancer_url)
> str(cancer_data)
$ age : Factor w/ 7 levels “20-29”,”30-39”,..: 7 3 4 4 3 3 4 4 3 3 ...
$ menopause : Factor w/ 4 levels “ge40”,”lt40”,..: 4 3 1 1 3 3 3 1 3 3 ...
$ tumor.size : Factor w/ 12 levels “0-4”,”10-14”,..: 12 3 3 7 7 6 5 8 2 1 ...

$ inv.nodes : Factor w/ 8 levels “0-2”,”12-14”,..: 8 1 1 1 1 5 5 1 1 1 ...
$ node.caps : Factor w/ 4 levels “”,”no”,”String”,..: 3 4 2 2 4 4 2 2 2 2 ...
$ deg.malig : Factor w/ 4 levels “1”,”2”,”3”,”String”: 4 3 1 2 3 2 2 3 2 2 ...
$ breast : Factor w/ 3 levels “left”,”right”,..: 3 2 2 1 2 1 2 1 1 2 ...
$ breast.quad: Factor w/ 7 levels “”,”central”,”left_low”,..: 7 4 2 3 3 6 4 4 4 5
$ irradiat : Factor w/ 3 levels “no”,”String”,..: 2 1 1 1 3 1 3 1 1 1 ...
$ Class : Factor w/ 3 levels “no-recurrence-events”,..: 3 2 1 2 1 2 1 1 1 1
> local_copy <- “cancer.csv”

> download.file(cancer_url, local_copy)
trying URL ‘http://repository.seasr.org/Datasets/UCI/csv/breast-cancer.csv’
Content type ‘application/octet-stream’ length 18804 bytes (18 KB)
downloaded 18 KB
> cancer_data <- read.csv(local_copy)
3.3. Accessing Databases

R can connect to all database management systems (DBMS) like SQLite, MySQL,
MariaDB, PostgreSQL and Oracle using the DBI package. We need to install and
load the DBI package and the backend package RSQLite. Define a database driver
of type SQLite using the function dbDriver() and setup a connection to the database
using the function dbConnect(). To retrieve data from the databases you write a
query as a string containing SQL commands and send it to the database with the
function dbGetQuery().
> install.packages(“DBI”)
> install.packages(“RSQLite”)
> library(DBI)
> library(RSQLite)
> driver <- dbDriver(“SQLite”)
91 Data Preparation
> db_file <- system.file(“extdata”, “crabtag.sqlite”, package = “learningr”)

> conn <- dbConnect(driver, db_file)
> query <- “SELECT * FROM IdBlock”
> id_block <- dbGetQuery(conn, query)
> id_block
Tag ID Firmware Version No Firmware Build Level
1 A03401 2 70
Alternatively, the function dbReadTable() reads a table from the connected

database and the function dbListTables() can list all the tables in the database.
> dbReadTable(conn, “idblock”)
Tag.ID Firmware.Version.No Firmware.Build.Level
1 A03401 2 70
> dbListTables(conn)
[1] “Daylog” “DeploymentNotebook” “IdBlock”
[4] “LifetimeNotebook” “TagNotebook”
The function dbDisconnect() is used for disconnecting and unloading the driver
and the function dbUnloadDriver() is used to unload the defined database driver.
> dbDisconnect(conn)
> dbUnloadDriver(driver)
For MySQL database we need to load the RMySQL package and set the
driver type to be “MySQL”. The PostgreSQL, Oracle and JDBC databases need
the PostgreSQL, ROracle and RJDBC packages respectively. To connect to an
SQL Server or Access databases, the RODBC package needs to be loaded. In this
package, the function odbcConnect() is used to connect to the database and the
function sqlQuery() is used to run a query and the function odbcClose() is used to
close and cleanup the database connections. There are not much matured methods
to access the NoSQL (Not only SQL) databases (lightweight databases – scalable
than traditional SQL relational databases). To access the MongoDB database the
packages RMongo and rmongodb are used. The database Cassandra can be accessed
using the package RCassandra.
3.4. Data Cleaning and Transforming
3.4.1. Manipulating Stings
In some datasets or data frames logical values are represented as “Y” and “N” instead
of TRUE and FALSE. In such cases it is possible to replace the string with correct
logical value as in the example below.
> a <- c(1,2,3)
> b <- c(“A”, “B”, “C”)
> d <- c(“Y”, “N”, “Y”)
> df1 <- data.frame(a, b, d)
> df1
a b d
1 1 A Y
2 2 B N
3 3 C Y
convt <- function(x)
{
y <- rep.int(NA, length(x))
y[x == “Y”] <- TRUE
y[x == “N”] <- FALSE
y
}
> df1$d <- convt(df1$d)
> df1
a b d
1 1 A TRUE
2 2 B FALSE
3 3 C TRUE
93 Data Preparation
The functions grep() and grepl() are used to find a pattern in a given text and the
functions sub() and gsub() are used to replace a pattern with another in a given text.
The above four functions belong to the base package, but the package stringr consists
of many such string manipulation functions. The function str_
detect() in the stringr package does the same function of detecting the presence of
a given pattern in the given text. We can also use the function fixed() to mention if
the string that we are searching for is a fixed one.
> grep(“my”, “This is my pen”)
[1] 1
> grepl(“my”, “This is my pen”)
[1] TRUE
> sub(“my”, “your”,”This is my pen”)
[1] “This is your pen”
> gsub(“my”, “your”,”This is my pen”)
[1] “This is your pen”
> str_detect(“This is my pen”, “my”)
[1] TRUE
> str_detect(“This is my pen”, fixed(“my”))
[1] TRUE
In the function str_detect(), it is possible to specify the search pattern with a

pipe symbol “|” to denote, that we need to find either of the two patterns specified.
That is we may be looking for the presence of “,” or “and” in the given text as shown
below.
> str_detect(“I like mangoes, oranges and pineapples”, “,|and”)
[1] TRUE
The function str_split() is used to split a given text based on the pattern specified
as below. This function returns a vector. But the function str_split_fixed() can be
used to split the given text into fixed number of strings based on the specified
patterns. This function returns a matrix.
> str_split(“I like mangoes, oranges and pineapples”, “,|and”)
[[1]]
[1] “I like mangoes” “ oranges “ “ pineapples”
> str_split_fixed(“I like mangoes, oranges and pineapples”, “,|and”, n = 3)
[,1] [,2] [,3]
[1,] “I like mangoes” “ oranges “ “ pineapples”
The function str_count() can be used to count the number of occurrence of a

given pattern in the given text.
> str_count(“I like mangoes, oranges and pineapples”, “a|o”)
[1] 6
> str_count(“I like mangoes, oranges and pineapples”, “s”)
[1] 3
The function str_replace() can be used to replace the specified pattern with
another pattern in the given text. This function will only replace the first occurrence
of the pattern. Hence, to replace all the occurrences of the pattern we use the
function str_replace_all(). In these functions, to denote multiple patterns to be
replaced, they can be placed within square brackets. This means it should replace
all that matches these characters specified within the square brackets.
> str_replace(“I like mangoes, oranges and pineapples”, “s”, “sss”)
[1] “I like mangoesss, oranges and pineapples”
> str_replace_all(“I like mangoes, oranges and pineapples”, “s”, “sss”)
[1] “I like mangoesss, orangesss and pineapplesss”
> str_replace_all(“I like mangoes, oranges and pineapples”, “[ao]”, “-”)
[1] “I like m-ng-es, -r-nges -nd pine-pples”
In the example below, the various ways of storing the gender values are
transformed into one way, ignoring the case differences. This is done using the
str_replace() function and the fixed() functions that ignores the case.
95 Data Preparation
> gender <- c(“MALE”, “Male”, “male”, “FEMALE”, “Female”, “female”)

> clean_gender <- str_replace(gender, fixed(“male”, ignore_case = TRUE), “Male”)
> clean_gender <- str_replace(clean_gender, fixed(“female”, ignore_case = TRUE),
male”)
> clean_gender
[1] “Male” “Male” “Male” “Female” “Female” “Female”
3.4.2. Manipulating Data Frames
To add a column to a data frame, we can use the below command to achieve this.
> name <- c(“Jhon”, “Peter”, “Mark”)
> start_date <- c(“1980-10-10”, “1999-12-12”, “1990-04-05”)
> end_date <- c(“1989-03-08”, “2004-09-20”, “2000-09-25”)
> service <- data.frame(name, start_date, end_date)
> service
name start_date end_date
1 Jhon 1980-10-10 1989-03-08
2 Peter 1999-12-12 2004-09-20
3 Mark 1990-04-05 2000-09-25
> service$period <- as.Date(service$end_date) - as.Date(service$start_date)
> service
name start_date end_date period
1 Jhon 1980-10-10 1989-03-08 3071 days
2 Peter 1999-12-12 2004-09-20 1744 days
3 Mark 1990-04-05 2000-09-25 3826 days
The same can be achieved using the function with() as below.

> service$period <- with(service, as.Date(end_date) - as.Date(start_date))
> service
name start_date end_date period

1 Jhon 1980-10-10 1989-03-08 3071 days
2 Peter 1999-12-12 2004-09-20 1744 days
3 Mark 1990-04-05 2000-09-25 3826 days
Another way of doing the same is using the function within(). But, the difference
lies when there are multiple columns to be added to a data frame, we can easily do
the same using the within() function in a single command and this is not possible
using the with() function.
> service <- within(service,
{
period <- as.Date(end_date) - as.Date(start_date)
highperiod <- period > 2000
})
> service
name start_date end_date period highperiod
1 Jhon 1980-10-10 1989-03-08 3071 days TRUE
2 Peter 1999-12-12 2004-09-20 1744 days FALSE
3 Mark 1990-04-05 2000-09-25 3826 days TRUE
The mutate() function in the plyr package also does the same function as the
function within(), but the syntax is slightly different.
> library(plyr)
> service <- mutate(service,
{
period = as.Date(end_date) - as.Date(start_date)
highperiod = period > 2000
})
> service
97 Data Preparation
name start_date end_date period highperiod

1 Jhon 1980-10-10 1989-03-08 3071 days TRUE
2 Peter 1999-12-12 2004-09-20 1744 days FALSE
3 Mark 1990-04-05 2000-09-25 3826 days TRUE
The function complete.cases() returns the number of rows in a data frame that
is free of missing values. The function na.omit() will remove the rows with missing
values in a data frame. And the function na.fail() throws an error message if the
data frame contains any missing values.
> crime.data <- read.csv(“F:/Crimes.csv”)
> nrow(crime.data)
[1] 65535
> complete <- complete.cases(crime.data)
> nrow(crime.data[complete, ])
[1] 63799
> clean.crime.data <- na.omit(crime.data)
> nrow(clean.crime.data)
[1] 63799
A data frame can be transformed by choosing few of the columns and ignoring
the remaining, but considering all the rows as in the example below.
> crime.data <- read.csv(“F:/Crimes.csv”)
> colnames(crime.data)
[1] “CASE.” “DATE..OF.OCCURRENCE” “BLOCK”
[4] “IUCR” “PRIMARY.DESCRIPTION”
“SECONDARY.DESCRIPTION”
[7] “LOCATION.DESCRIPTION” “ARREST” “DOMESTIC”
[10] “BEAT” “WARD” “FBI.CD”
[13] “X.COORDINATE” “Y.COORDINATE” “LATITUDE”
[16] “LONGITUDE” “LOCATION”
> crime.data1 <- crime.data[, 1:6]

> colnames(crime.data1)
[1] “CASE.” “DATE..OF.OCCURRENCE” “BLOCK”
[4] “IUCR” “PRIMARY.DESCRIPTION”
“SECONDARY.DESCRIPTION”
Alternatively, the data frame can be transformed by selecting only the required
rows and retaining all columns of a data frame as in the example below.
> nrow(crime.data)
[1] 65535
> crime.data2 <- crime.data[1:10,]
> nrow(crime.data2)
[1] 10
The function sort() sorts the given vector of numbers or strings. It generally
sorts from smallest to largest, but this can be altered using the argument decreasing
= TRUE.
> x <- c(5, 10, 3, 15, 6, 8)
> sort(x)
[1] 3 5 6 8 10 15
> sort(x, decreasing = TRUE)
[1] 15 10 8 6 5 3
> y <- c(“X”, “AB”, “Deer”, “For”, “Moon”)
> sort(y)
[1] “AB” “Deer” “For” “Moon” “X”
> sort(y, decreasing = TRUE)
[1] “X” “Moon” “For” “Deer” “AB”
The function order() is the inverse of the sort() function. It returns the index
of the vector elements in the order as below. But, x[order(x)] is same as sort(x). This
can be seen by the use of the identical() function.
99 Data Preparation
> order(x)
[1] 3 1 5 6 2 4
> x[order(x)]
[1] 3 5 6 8 10 15
> identical(sort(x), x[order(x)])
[1] TRUE
The order() function is more useful than the sort() function as it can be used to
manipulate the data frames easily.
> name <- c(“Jhon”, “Peter”, “Mark”)
> start_date <- c(“1980-10-10”, “1999-12-12”, “1990-04-05”)
> end_date <- c(“1989-03-08”, “2004-09-20”, “2000-09-25”)
> service <- data.frame(name, start_date, end_date)
> service
1 Jhon 1980-10-10 1989-03-08
2 Peter 1999-12-12 2004-09-20
3 Mark 1990-04-05 2000-09-25
> startdt <- order(service$start_date)
> service.ordered <- service[startdt, ]
> service.ordered
1 Jhon 1980-10-10 1989-03-08
3 Mark 1990-04-05 2000-09-25
2 Peter 1999-12-12 2004-09-20
The arrange() function of the plyr package does the same function as above.
> library(plyr)
> arrange(service, start_date)

1 Jhon 1980-10-10 1989-03-08
2 Mark 1990-04-05 2000-09-25
3 Peter 1999-12-12 2004-09-20
The rank() function lists the rank of the elements in a vector or a data frame.
By specifying the argument ties.method = “first”, a rank need not be shared among
more than one element with the same value.
> x <- c(9, 5, 4, 6, 4, 5)
> rank(x)
[1] 6.0 3.5 1.5 5.0 1.5 3.5
> rank(x, ties.method = “first”)
[1] 6 3 1 5 2 4
The SQL statements can be executed from R and the results can be obtained
as in any other database. The package sqldf needs to be installed to manipulate the
data frames or datasets using SQL.
> install.packages(“sqldf ”)
> library(sqldf)
> query <- “SELECT * FROM iris WHERE Species = ‘setosa’”
> sqldf(query)
3.4.3. Data Reshaping
Data Reshaping in R is about changing the way data is organized into rows and
columns. Most of the time data processing in R is done by taking the input data as a
data frame. It is easy to extract data from the rows and columns of a data frame. But
there are situations when we need the data frame in a different format than what we
received. R has few functions to split, merge and change the columns to rows and vice-
versa in a data frame.
The cbind() function can be used to join multiple vectors to create a data frame.
We can also merge two data frames using the rbind() function.
101 Data Preparation
> city <- c(“Tampa”,”Seattle”,”Hartford”,”Denver”)

> state <- c(“FL”,”WA”,”CT”,”CO”)
> zipcode <- c(33602,98104,06161,80294)
> addresses <- cbind(city,state,zipcode)
> addresses
city state zipcode
[1,] “Tampa” “FL” “33602”
[2,] “Seattle” “WA” “98104”
[3,] “Hartford” “CT” “6161”
[4,] “Denver” “CO” “80294”
> new.address <- data.frame(
+ city = c(“Lowry”,”Charlotte”),
+ state = c(“CO”,”FL”),
+ zipcode = c(“80230”,”33949”),
+ stringsAsFactors = FALSE
+)
> print(new.address)
city state zipcode
1 Lowry CO 80230
2 Charlotte FL 33949
> all.addresses <- rbind(addresses,new.address)
> all.addresses
city state zipcode
1 Tampa FL 33602
2 Seattle WA 98104
3 Hartford CT 6161
4 Denver CO 80294
5 Lowry CO 80230
6 Charlotte FL 33949
The merge() function can be used to merge two data frames. The merging
requires the data frames to have same column names on which the merging is done.
In the example below, we consider the data sets about Diabetes in Pima Indian
Women available in the library named “MASS”. The two datasets are merged based
on the values of blood pressure (“bp”) and body mass index (“bmi”). On choosing
these two columns for merging, the records where values of these two variables
match in both data sets are combined together to form a single data frame.
> library(MASS)
> head(Pima.te)
npreg glu bp skin bmi ped age type
1 6 148 72 35 33.6 0.627 50 Yes
2 1 85 66 29 26.6 0.351 31 No
3 1 89 66 23 28.1 0.167 21 No
...
> head(Pima.tr)
npreg glu bp skin bmi ped age type
1 5 86 68 28 30.2 0.364 24 No
2 7 195 70 33 25.1 0.163 55 Yes
3 5 77 82 41 35.8 0.156 35 No
...
> nrow(Pima.te)
[1] 332
> nrow(Pima.tr)
[1] 200
> merged.Pima <- merge(x = Pima.te, y = Pima.tr,
+ by.x = c(“bp”, “bmi”),
+ by.y = c(“bp”, “bmi”)
+)
> head(merged.Pima)
bp bmi npreg.x glu.x skin.x ped.x age.x type.x npreg.y glu.y skin.y
1 60 33.8 1 117 23 0.466 27 No 2 125 20
2 64 29.7 2 75 24 0.370 33 No 2 100 23
3 64 31.2 5 189 33 0.583 29 Yes 3 158 13
...
ped.y age.y type.y
1 0.088 31 No
2 0.368 21 No
3 0.295 24 No
...
> nrow(merged.Pima)
[1] 17
One of the most interesting aspects of R programming is about changing the

shape of the data in multiple steps to get a desired shape. The functions used to do
this are called melt() and cast(). We consider the dataset called ships present in the
library called “MASS”.
> library(MASS)
> head(ships)
type year period service incidents
1 A 60 60 127 0
2 A 60 75 63 0
3 A 65 60 1095 3
...
Now we melt the data using the melt() function in the package reshape2 to
organize it, converting all columns other than type and year into multiple rows.
> library(reshape2)
> molten.ships <- melt(ships, id = c(“type”,”year”))
> head(molten.ships)
type year variable value
1 A 60 period 60
2 A 60 period 75
3 A 65 period 60
4 A 65 period 75
5 A 70 period 60
6 A 70 period 75
> nrow(molten.ships)
[1] 120
> nrow(ships)
[1] 40
We can cast the molten data into a new form where the aggregate of each type
of ship for each year is created. It is done using the cast() function.
> recasted.ship <- cast(molten.ships, type+year~variable,sum)
> head(recasted.ship)
type year period service incidents
1 A 60 135 190 0
2 A 65 135 2190 7
3 A 70 135 4865 24
4 A 75 135 2244 11
5 B 60 135 62058 68
6 B 65 135 48979 111
3.4.4. Grouping Functions
R has many apply functions such as apply(), lapply(), sapply(), vapply(), mapply(),
rapply(), tapply(), aggregate() and by(). Function lapply() is a list apply which acts
on a list or vector and returns a list. Function sapply() is a simple lapply() function
defaults to returning a vector or matrix when possible. Function vapply() is a verified
apply() function that allows the return object type to be pre-specified. Function
rapply() is a recursive apply for nested lists, i.e. lists within lists. Function tapply()
is a tagged apply where the tags identify the subsets. Function apply() is generic,
applies a function to a matrix’s rows or columns or, more generally, to dimensions
of an array.
If we want to apply a function to the rows or columns of a matrix or array, we

use the apply() function as below.
> M <- matrix(seq(1,16), 4, 4)
>M
[,1] [,2] [,3] [,4]
[1,] 1 5 9 13
[2,] 2 6 10 14
[3,] 3 7 11 15
[4,] 4 8 12 16
> apply(M, 1, min)

[1] 1 2 3 4
> apply(M, 2, max)

[1] 4 8 12 16
> M <- array( seq(32), dim = c(4,4,2))
>M
,,1
[,1] [,2] [,3] [,4]

[1,] 1 5 9 13
[2,] 2 6 10 14
[3,] 3 7 11 15
[4,] 4 8 12 16
,,2
[,1] [,2] [,3] [,4]
[1,] 17 21 25 29
[2,] 18 22 26 30
[3,] 19 23 27 31
[4,] 20 24 28 32
> apply(M, 1, sum)

[1] 120 128 136 144
> apply(M, c(1,2), sum)

[,1] [,2] [,3] [,4]
[1,] 18 26 34 42
[2,] 20 28 36 44
[3,] 22 30 38 46
[4,] 24 32 40 48
If we want to apply a function to each element of a list in turn and get a list
back, we use the lapply() function as below.
> x <- list(a = 1, b = 1:3, c = 10:100)
>x
$a
[1] 1
$b
[1] 1 2 3
$c
[1] 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26
[18] 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43
[35] 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
[52] 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77
[69] 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94
[86] 95 96 97 98 99 100
> lapply(x, FUN = length)

$a
[1] 1
$b
[1] 3
$c
[1] 91
> lapply(x, FUN = sum)

$a
[1] 1
$b
[1] 6
$c
[1] 5005
We use the function sapply(), if we want to apply a function to each element of

a list in turn, and we want a vector back.
> x <- list(a = 1, b = 1:3, c = 10:100)
>x
$a
[1] 1
$b
[1] 1 2 3
$c
[1] 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26
[18] 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43
[35] 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
[52] 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77
[69] 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94
[86] 95 96 97 98 99 100
> sapply(x, FUN = length)

a b c
1 3 91
> sapply(x, FUN = sum)

a b c
1 6 5005
When we want to use the function sapply(), but need to squeeze some more speed
out of the code, we use the function vapply() as below. For the function vapply(), we
give R the information on what the function will return, which can save some time
coercing returned values to fit in a single atomic vector. In the example below, we tell
R that everything returned by length() should be an integer of length 1.
> x <- list(a = 1, b = 1:3, c = 10:100)
> vapply(x, FUN = length, FUN.VALUE = 0L)
a b c
1 3 91
For when we have several data structures (e.g. vectors, lists) and we want to
apply a function to the 1st elements of each, and then the 2nd elements of each,
etc., coercing the result to a vector/array we use the function vapply() as below.
> mapply(sum, 1:5, 1:5, 1:5)

[1] 3 6 9 12 15
> mapply(rep, 1:4, 4:1)
[[1]]
[1] 1 1 1 1
[[2]]
[1] 2 2 2
[[3]]
[1] 3 3
[[4]]
[1] 4
When we want to apply a function to each element of a nested list structure,

recursively, we use the function rapply() as below. The function rapply() can be best
illustrated with a user defined function to be applied.
> myFun <- function(x){
+ if (is.character(x)){
+ return(paste(x,”!”,sep=””))
+ }
+ else{
+ return(x + 1)
+ }
+}
> l <- list(a = list(a1 = “Boo”, b1 = 2, c1 = “Eeek”),

+ b = 3, c = “Yikes”,
+ d = list(a2 = 1, b2 = list(a3 = “Hey”, b3 = 5)))
> rapply(l, myFun)

a.a1 a.b1 a.c1 b c d.a2 d.b2.a3 d.b2.b3

“Boo!” “3” “Eeek!” “4” “Yikes!” “2” “Hey!” “6”
> rapply(l, myFun, how = “replace”)
$a
$a$a1
[1] “Boo!”
$a$b1
[1] 3
$a$c1
[1] “Eeek!”
$b
[1] 4
$c
[1] “Yikes!”
$d
$d$a2
[1] 2
$d$b2
$d$b2$a3
[1] “Hey!”
$d$b2$b3
[1] 6
When we want to apply a function to subsets of a vector and the subsets are
defined by some other vector, usually a factor, we use the function tapply() as below.
> x <- 1:20
>x
[1] 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20
> y <- factor(rep(letters[1:5], each = 4))

>y
[1] a a a a b b b b c c c c d d d d e e e e
Levels: a b c d e
> tapply(x, y, sum)

a b c d e
10 26 42 58 74
The by() function, can be thought of, as a “wrapper” for the function tapply().
When we want to compute a task that tapply() can’t handle, the by() function
arises.
> cta <- tapply(iris$Sepal.Width , iris$Species , summary )
> cba <- by(iris$Sepal.Width , iris$Species , summary )
> cta
$setosa
2.300 3.200 3.400 3.428 3.675 4.400
$versicolor
2.000 2.525 2.800 2.770 3.000 3.400
$virginica
2.200 2.800 3.000 2.974 3.175 3.800
> cba
iris$Species: setosa
2.300 3.200 3.400 3.428 3.675 4.400
------------------------------------------------------
iris$Species: versicolor
2.000 2.525 2.800 2.770 3.000 3.400
------------------------------------------------------
iris$Species: virginica
2.200 2.800 3.000 2.974 3.175 3.800
If we print these two objects, cta and cba, we have the same results. The only
differences are in how they are shown with the different class attributes. The power
of the function by() arises when we can’t use the function tapply() as in the following
code.
> tapply(iris, iris$Species, summary )
Error in tapply(iris, iris$Species, summary) :
arguments must have same length
R says that arguments must have the same lengths, say “we want to calculate
the summary of all variable in iris along the factor Species”: but R just can’t do
that because it does not know how to handle. The by() function lets the summary()
function work even if the length of the first argument are different.
> bywork <- by(iris, iris$Species, summary )
> bywork
iris$Species: setosa
Sepal.Length Sepal.Width Petal.Length Petal.Width
Min. :4.300 Min. :2.300 Min. :1.000 Min. :0.100
1st Qu.:4.800 1st Qu.:3.200 1st Qu.:1.400 1st Qu.:0.200
Median :5.000 Median :3.400 Median :1.500 Median :0.200
Mean :5.006 Mean :3.428 Mean :1.462 Mean :0.246
3rd Qu.:5.200 3rd Qu.:3.675 3rd Qu.:1.575 3rd Qu.:0.300
Max. :5.800 Max. :4.400 Max. :1.900 Max. :0.600
Species
setosa :50
versicolor: 0
virginica : 0
------------------------------------------------------
iris$Species: versicolor
Min. :4.900 Min. :2.000 Min. :3.00 Min. :1.000
Max. :7.000 Max. :3.400 Max. :5.10 Max. :1.800
Species
setosa :0
versicolor:50
virginica : 0
------------------------------------------------------
iris$Species: virginica
Min. :4.900 Min. :2.200 Min. :4.500 Min. :1.400
Max. :7.900 Max. :3.800 Max. :6.900 Max. :2.500
Species
setosa :0
versicolor: 0
virginica :50
The arguments must have the same lengths. R can’t do that because it does not
know how to handle it. The by() function lets the summary() function work even if
the length of the first argument is different. The result is an object of class by that
along Species computes the summary of each variable.
The aggregate() function can be seen as another a different way of using tapply()
function if we use it in such a way.
> att <- tapply(iris$Sepal.Length , iris$Species , mean)
> agt <- aggregate(iris$Sepal.Length , list(iris$Species), mean)
> att
setosa versicolor virginica
5.006 5.936 6.588
> agt
Group.1 x
1 setosa 5.006
2 versicolor 5.936
3 virginica 6.588
The two immediate differences are that the second argument of the aggregate()
function must be a list while tapply() function can (not mandatory) be a list and
that the output of the aggregate() function is a data frame while the one of tapply()
function is an array. The power of the aggregate() function is that it can handle
easily subsets of the data with subset argument and that it can handle formula
as well. These elements make the aggregate() function easier to work with than
tapply() function in some situations.
> ag <- aggregate(len ~ ., data = ToothGrowth, mean)
> ag
supp dose len

1 OJ 0.5 13.23
2 VC 0.5 7.98
3 OJ 1.0 22.70
4 VC 1.0 16.77
5 OJ 2.0 26.06
6 VC 2.0 26.14
 HIGHLIGHTS
 One of the packages in R is datasets which is filled with example datasets.
 We can see all the datasets available in the loaded packages using the
data() function.
 The read.table() function reads the CSV files and stores the result in a
data frame.
 The system.file() function is used to locate files that are inside a package.
 Writing data from R into a file is done using the functions write.table()
and write.csv().
 If the file is unstructured, it is read using the function readLines().
 The writeLines() function takes a text line and the file name as argument
and writes the text to the file.
 The XML file can be imported using the function xmlParse() function.
 The function used to import the JSON file is fromJSON() and the function
used to export the JSON file is toJSON().
 Spreadsheets can be imported with the functions read.xlsx() and read.
xlsx2().
 To write to an excel file from R we use the function write.xlsx2().
 The read.ssd() function is used to read SAS datasets.
 The read.spss() function is used to import the SPSS data files.
 The MATLAB binary data files can be read and written using the readMat()
and writeMat() functions in the R.matlab package.
 R can connect to all DBMS like SQLite, MySQL, MariaDB, PostgreSQL

and Oracle using the DBI package.
 The function dbReadTable() reads a table from the connected database.
 The functions grep() and grepl() are used to find a pattern in a given text.
 The functions sub() and gsub() are used to replace a pattern with another
in a given text.
 The function str_split() is used to split a given text based on the pattern
specified.
 The function str_count() can be used to count the number of occurrence
of a given pattern in the given text.
 The function str_replace() can be used to replace the specified pattern
with another pattern in the given text.
 The function na.omit() will remove the rows with missing values in a data
frame.
 The function sort() sorts the given vector of numbers or strings.
 The function order() is the inverse of the sort() function.
 The rank() function lists the rank of the elements in a vector or a data
frame.
 The package sqldf needs to be installed to manipulate the data frames or
datasets using SQL.
 Changing the shape of a data frame is done using the functions melt() and
cast().
 R has many grouping functions such as apply(), lapply(), sapply(), vapply(),
mapply(), rapply(), tapply(), aggregate() and by().
CHAPTER 4
Graphics using R
 OBJECTIVES

 understand what is Exploratory Data Analysis
 introduce about the main graphical packages
 draw pie charts using R
 draw scatter plots using R
 draw line plots using R
 draw histograms using R
 draw box plots using R
 draw bar plots using R
 know about other existing graphical packages
4.1. Exploratory Data Analysis

Exploratory Data Analysis (EDA) is a visual based method used to analyse data
sets and to summarize their main characteristics. Exploratory Data Analysis (EDA)
shows how to use visualisation and transformation to explore data in a systematic
way. EDA is an iterative cycle of the below steps:
1) Generate questions about data.
2) Search for answers by visualising, transforming, and modelling data.
3) Use what is learnt to refine questions and/or generate new questions.
Exploratory Data Analysis (EDA) is an approach for data analysis that employs
a variety of techniques (mostly graphical) to:
1) Maximize insight into a data set
2) Uncover underlying structure
3) Extract important variables
4) Detect outliers and anomalies
5) Test underlying assumptions
6) Develop parsimonious models
7) Determine optimal factor settings.
4.2. Main Graphical Packages

The basic graphs in R can be drawn using the base graphics system. These have
some limitations and they are overcome in the next level of graphics called the grid
graphics system. This system allows to plot the points or lines in the place where
desired. But, this does not allow us to draw a scatter plot. Hence, we go for the next
level of plotting which the lattice graphics system is. In this system, the results of
a plot can be saved. Also these scatter plots can contain multiple panels in which
we can draw multiple graphs and compare them to each other. The next levels of
graphs are the ggplot2 graphics system. In this the “gg” stands for “grammar of
graphics”. This breaks down the graphs into many parts or chunks.
4.3. Pie Charts

In R the pie chart is created using the pie() function which takes positive numbers
as vector input. The additional parameters are used to control labels, colour, title
etc. The basic syntax for creating a pie-chart is as given below and the explanation
of the parameters are also listed.
pie(x, labels, radius, main, col, clockwise)
x – numeric vector
labels – description of the slices
radius – values between [-1 to +1]
119 Graphics using R
main – title of the chart

col – colour palette
clockwise – logical value – TRUE (Clockwise), FALSE (Anti Clockwise)
> x <- c(25, 35, 10, 5, 15)

> labels <- c(“Rose”, “Lotus”, “Lilly”, “Sunflower”, “Jasmine”)
> pie(x, labels = percent, main = “Flowers”, col = rainbow(length(x)))
> legend(“topright”, c(“Rose”, “Lotus”, “Lilly”, “Sunflower”, “Jasmine”),
cex = 0.8, fill = rainbow(length(x)))
Figure 4.1 Pie Chart of Flowers
A 3D Pie Chart can be drawn using the package plotrix which uses the function
pie3D().
> install.packages(“plotrix”)
> library(plotrix)
> pie3D(x, labels = labels, explode = 0.1, main = “Flowers”)
Figure 4.2 3-D Pie Chart of Flowers
4.4. Scatter Plots

Scatter plots are used for exploring the relationship between the two continuous
variables. Let us consider the dataset “cars” that lists the “Speed and Stopping
Distances of Cars”. The basic scatter plot in the base graphics system can be
obtained by using the plot() function as in Fig. 4.3. The below example compares if
the speed of a car has effect on its stopping distance using the plot.
> colnames(cars)
[1] “speed” “dist”
> plot(cars$speed, cars$dist)
Figure 4.3 Basic Scatter Plot of Car Speed Vs Distance
This plot can be made more appealing and readable by adding colour and
changing the plotting character. For this we use the arguments col and pch (can
take the values between 1 and 25) in the plot() function as below. Thus the plot in
Fig. 4.4 shows that there is a strong positive correlation between the speed of a car
and its stopping distance.
> plot(cars$speed, cars$dist, col = “red”, pch = 15)
Figure 4.4 Coloured Scatter Plot of Car Speed Vs Distance

The layout() function is used to control the layout of multiple plots in the
matrix. Thus in the example below multiple related plots are placed in a single
figure as in Fig. 4.5.
> data(mtcars)
> layout(matrix(c(1,2,3,4), 2, 2, byrow = TRUE))
> plot(mtcars$wt, mtcars$mpg, col = “blue”, pch = 17)
> plot(mtcars$wt, mtcars$disp, col = “red”, pch = 15)
> plot(mtcars$mpg, mtcars$disp, col = “dark green”, pch = 10)
> plot(mtcars$mpg, mtcars$hp, col = “violet”, pch = 7)
Figure 4.5 Layout of Multiple Scatter Plots
When we have more than two variables and we want to find the correlation
between one variable versus the remaining ones we use scatter plot matrix. We use
pairs() function to create matrices of scatter plots as in Fig. 4.6. The basic syntax for
creating scatter plot matrices in R is as below.
pairs(formula, data)
> pairs(~wt+mpg+disp+cyl,data = mtcars, main = “Scatterplot Matrix”)

Figure 4.6 Scatter Plot Matrix Using pairs()
The lattice graphics system has equivalent of plot() function and it is xyplot().
This function uses a formula to specify the x and y variables (yvar ~ xvar) and a data
frame argument. To use this function, it is required to include the lattice package.
> library(lattice)
> xyplot(mtcars$mpg ~ mtcars$disp, mtcars, col = “purple”, pch = 7)
Figure 4.7 Scatter Plot Matrix Using xyplot()

Axis scales can be specified in the xyplot() using the scales argument and this
argument must be a list. This list consists of the name = value pairs. If we mention
log = TRUE, the log scales for the x and y axis are set as in Fig. 4.8. The scales list
can take other arguments also like the x and y that sets the x and y axes respectively.
> xyplot(mtcars$mpg ~ mtcars$disp, mtcars, scales = list(log = TRUE),
col = “red”, pch = 11)
Figure 4.8 Scatter Plot Matrix with Axis Scales Using xyplot()
The data in the graph can be split based on one of the columns in the dataset
namely mtcars$carb. This can be done by appending the pipe symbol (|) along with
the column name used for splitting. The argument relation = “same” means that
each panel shares the same axes. If the argument alternating = TRUE, axis ticks for
each panel is drawn on alternating sides of the plot as in Fig. 4.9.
> xyplot(mtcars$mpg ~ mtcars$disp | mtcars$carb, mtcars,
scales = list(log = TRUE, relation = “same”, alternating = FALSE),
layout = c(3, 2), col = “blue”, pch = 14)
Figure 4.9 Scatter Plot Split on a Column
The lattice plots can be stored in variables and hence they can be further
updated using the function update as below.
> graph1 <- xyplot(mtcars$mpg ~ mtcars$disp | mtcars$carb, mtcars,
scales = list(log = TRUE, relation = “same”, alternating = FALSE),
layout = c(3, 2), col = “blue”, pch = 14)
> graph2 <- update(graph1, col = “yellow”, pch = 6)
In the ggplot2 graphics, each plot is drawn with a call to the ggplot() function
as in Fig. 4.10. This function takes a data frame as its first argument. The passing
of data frame columns to the x and y axis is done using the aes() function which is
used within the ggplot() function. The other aesthetics to the graph are then added
using the geom() function appended with a “+” symbol to the ggplot() function.
> library(ggplot2)
> ggplot(mtcars, aes(mpg, disp)) +
geom_point(color = “purple”, shape = 16, cex = 2.5)
Figure 4.10 Scatter Plot Using ggplot()
The ggplots can also be split into several panels like the lattice plots as in Fig. 4.11.
This is done using the function facet_wrap() which takes a formula of the column
used for splitting. The function theme() is used to specify the orientation of the
axis readings. The functions facet_wrap() and theme() are appended to the ggplot()
function using the “+” symbol. The ggplots can be stored in a variable like the lattice
plots and as usual wrapping the expression in parentheses makes it to auto print.
> (graph1 <- ggplot(mtcars, aes(mpg, disp)) +
geom_point(color = “dark green”, shape = 15, cex = 3))
> (graph2 <- graph1 + facet_wrap(~mtcars$cyl, ncol = 3) +
theme(axis.text.x = element_text(angle = 90, hjust = 1)
Figure 4.11 – Scatter Plot Split into Panels Using ggplot()

4.5. Line Plots

A line chart / line plot is a graph that connects a series of points by drawing line
segments between them. Line charts are usually used in identifying the trends in
data. The plot() function in R is used to create the line graph in base graphics as in
Fig. 4.12. This function takes a vector of numbers as input together with few more
parameters listed below.
plot(v, type, col, xlab, ylab)
v – numeric vector
type - takes value “p” (only points), or “l” (only lines) or “o” (both points and lines)
xlab – label of x-axis
ylab – label of y-axis
main - title of the chart
col - colour palette
Figure 4.12 Line Plot Using Basic Graphics

> male <- c(1000, 2000, 1500, 4000, 800)
> female <- c(700, 300, 600, 1200, 800)
> child <- c(1000, 1200, 1500, 800, 2000)
> wages <- c(“Male”, “Female”, “Children”)
> color = c(“red”, “blue”, “green”)
> plot(male, type = “o”, col = “red”, xlab = “Month”, ylab = “Wages”,
main = “Monthly Wages”, ylim = c(0, 5000))
> lines(female, type = “o”, col = “blue”)
> lines(child, type = “o”, col = “green”)
> legend(“topleft”, wages, cex = 0.8, fill = color)
Line plots in the lattice graphics uses the xyplot() function as in Fig. 4.13. In
this multiple lines can be creating using the “+” symbol in the formula where the
x and the y axes are mentioned. The argument type = “l” is used to mention that
it is a continuous line.
> xyplot(economics$pop + economics$unemploy ~ economics$date, economics, type = “l”)
Figure 4.13 Line Plot Using Lattice Graphics
In the ggplot2 graphics, the same syntax for scatter plots are used, except for
the change of geom_plot() function with the geom_line() function as in Fig. 4.14.
But, there need to be multiple geom_line() functions for multiple lines to be drawn
in the graph.
> ggplot(economics, aes(economics$date)) + geom_line(aes(y = economics$pop)) +
geom_line(aes(y = economics$unemploy))
Figure 4.14 Line Plot Using ggplot2 Graphics
The plot in the Fig. 4.15 can be drawn without using multiple geom_line()
functions also. This is possible using the function geom_ribbon() as mentioned
below. This function plots not only the two lines, but also the contents in between
the two lines.
> ggplot(economics, aes(economics$date, ymin = economics$unemploy,
ymax = economics$pop)) + geom_ribbon(color = “blue”, fill = “white”)
Figure 4.15 Line Plot Using geom_ribbon()

4.6. Histograms
Histograms represents the variable values frequencies, that are split into ranges. This
is similar to bar charts, but histograms group values into continuous ranges. In R
histograms in the base graphics are drawn using the function hist() as in the Fig. 4.16,
that takes a vector of numbers as input together with few more parameters listed below.
hist(v, main, xlab, xlim, ylim, breaks, col, border)
v – numeric vector main - title of the chart
col - colour palette border – border colour
xlab – label of x-axis xlim – range of x-axis
ylim – range of y-axis breaks – width of each bar
Figure 4.16 Histogram Using Base Graphics

> x <- c(45, 33, 31, 23, 58, 47, 39, 58, 28, 55, 42, 27)
> hist(x, xlab = “Age”, col = “blue”, border = “red”, xlim = c(25, 60),
ylim = c(0, 3), breaks = 5)
The lattice histogram is drawn using the function histogram() as in Fig. 4.17 and
it behaves in the same way as the base ones. But it allows easy splitting of data into
panels and saving plots as variables. The breaks argument behaves the same way as with
hist(). The lattice histograms support counts, probability densities, and percentage
y-axes via the type argument, which takes the string “count”, “density”, or “percent”.
> histogram(~ mtcars$mpg, mtcars, breaks = 10)
Figure 4.17 Histogram Using Lattice Graphics
The ggplot histograms are created by adding the function geom_histogram() to

the ggplot() function as in Fig. 4.18. Bin specification is simple here, we just need
to pass a numeric bin width to geom_histogram() function. It is possible to choose
between counts and densities by passing the special names ..count.. or ..density.. to
the y-aesthetic.
> ggplot(mtcars, aes(mtcars$mpg, ..density..)) + geom_histogram(binwidth = 5)
Figure 4.18 Histogram Using ggplot2 Graphics

4.7. Box Plots

The box plot divides the data into three quartiles. This graph represents the
minimum, maximum, median, first quartile and third quartile in the data. This
shows the data distribution by drawing the box plots. In R base graphics the box
plot is created using the boxplot() function as in Fig. 4.19, which takes the following
parameters. The parameters are used to give the data as a data frame, a vector or
a formula, a logical value to draw a notch, a logical value to draw a box as per the
width of the sample, give title of the chart, labels for the boxes. The basic syntax
for creating a box-plot is as given below and the explanation of the parameters are
also listed.
boxplot(x, data, notch, varwidth, names, main)
x – vector or a formula
data – data frame
notch – logical value (TRUE – draw a notch)
varwidth – logical value (TRUE – box width proportionate to sample size
names – labels printed under the boxes
main – title of the chart
Figure 4.19 Box Plots Using Base Graphics

This type of plot is often clearer if we reorder the box plots from smallest to
largest, in some sense. The reorder() function changes the order of a factor’s levels,
based upon some numeric score.
> boxplot(mpg ~ reorder(gear, mpg, median), data = mtcars,
xlab = “Number of Gears”, ylab = “Miles Per Gallon”,
main = “Car Mileage”, varwidth = TRUE,
col = c(“red”,”blue”, “green”), names = c(“Low”, “Medium”, “High”))
In the lattice graphics the box plot is drawn using the function bwplot() as in
Fig. 4.20.
> bwplot(mpg ~ reorder(gear, mpg, median), data = mtcars,
xlab = “Number of Gears”, ylab = “Miles Per Gallon”,
main = “Car Mileage”, varwidth = TRUE,
col = c(“red”,”blue”, “green”), names = c(“Low”, “Medium”, “High”))
Figure 4.20 Box Plots Using Lattice Graphics
In the ggplot2 graphics the box plot is drawn by adding the function geom_
boxplot() to the function ggplot() as in Fig. 4.21.
> ggplot(mtcars, aes(reorder(gear, mpg, median), mpg)) + geom_boxplot()
Figure 4.21 Box Plots Using ggplot2 Graphics
4.8. Bar Plots

Bar charts are the natural way of displaying numeric variables split by a categorical
variable. In R base graphics the bar chart is created using the barplot() function
as in Fig. 4.22, which takes a matrix or a vector of numeric values. The additional
parameters are used to give labels to the X-axis, Y-axis, give title of the chart, labels
for the bars and colours. The basic syntax for creating a bar-chart is as given below
and the explanation of the parameters are also listed.
barplot(H, xlab, ylab, main, names.arg, col)
H – numeric vector or matrix
x-lab – label of x-axis
y-lab – label of y-axis
main - title of the chart
names.arg – vector of labels under each bar
col – colour palette
> x <- matrix(c(1000, 900, 1500, 4400, 800, 2100, 1700, 2900, 3800),
nrow = 3, ncol = 3)
> years <- c(“2011”, “2012”, “2013”)
> city <- c(“Chennai”, “Mumbai”, “Kolkata”)
> color <- c(“red”, “blue”, “green”)

> barplot(x, main = “Yearly Sales”, names.arg = years, xlab = “Year”,
ylab = “Sales”, col = color)
> legend(“topleft”, city, cex = 0.8, fill = color)
Figure 4.22 Vertical Bar Plot Using Base Graphics
By default the bars are vertical, but if we want horizontal bars, they can be
generated with horiz = TRUE parameter as in Fig. 4.23. We can also do some
fiddling with the plot parameters, via the par() function. The las parameter controls
whether labels are horizontal, vertical, parallel, or perpendicular to the axes. Plots
are usually more readable if you set las = 1, for horizontal. The mar parameter is a
numeric vector of length 4, giving the width of the plot margins at the bottom/left/
top/right of the plot.
> x <- matrix(c(1000, 900, 1500, 4400, 800, 2100, 1700, 2900, 3800), nrow = 3, ncol = 3)
> years <- c(“2011”, “2012”, “2013”)
Figure 4.23 Horizontal Bar Plot Using Base Graphics

> city <- c(“Chennai”, “Mumbai”, “Kolkata”)
> color <- c(“red”, “blue”, “green”)
> par(las = 1, mar = c(3, 9, 1, 1))
> barplot(x, main = “Yearly Sales”, names.arg = years,
xlab = “Year”, ylab = “Sales”, col = color, horiz = TRUE)
> legend(“bottomright”, city, cex = 0.8, fill = color)
The lattice equivalent of the function barplot(), is the function barchart() as

shown in Fig. 4.24. The formula interface is the same as those we saw with scatter
plots, yvar ~ xvar.
> barchart(mtcars$mpg ~ mtcars$disp, mtcars)
Figure 4.24 Horizontal Bar Plot Using Lattice Graphics
Extending this to multiple variables just requires a tweak to the formula, and
passing stack = TRUE to make a stacked plot as in Fig. 4.25.
> barchart(mtcars$mpg ~ mtcars$disp + mtcars$qsec + mtcars$hp, mtcars,
stack = TRUE
Figure 4.25 Horizontal Stacked Bar Plot Using Lattice Graphics

In the ggplot2 graphics the bar chart is drawn by adding the function geom_
bar() to the function ggplot() as in Fig. 4.26. Like base, ggplot2 defaults to vertical
bars; adding the function coord_flip() swaps this. We must pass the argument stat
= “identity” to the function geom_bar().
> ggplot(mtcars, aes(mtcars$mpg, mtcars$disp)) + geom_bar(stat = “identity”) +
coord_flip()
Figure 4.26 Horizontal Bar Plot Using ggplot2 Graphics
4.9. Other Graphical Packages
Package Name Description

Plots for visualizing categorical data, such as mosaic plots
Vcd
and association plots
Plotrix Loads of extra plot types
latticeExtra Extends the lattice package
GGally Extend the ggplot2 package
Provide access to the underlying framework of lattice and
Grid
ggplot2 packages
Package Name Description

gridSVG Write grid-based plots to SVG files
Allows pointing and clicking to interact with base or lattice
Playwith
plots
Provides a whole extra system of plots with more
Iplots
interactivity
Provides an R wrapper around Google Chart Tools, creating
googleVis
plots that can be displayed in a browser.
Provides an interface to GGobi package (for visualizing
Rggobi
high-dimensional data)
Open source visualization program for exploring high-
GGobi
dimensional data.
Rgl Provides an interface to OpenGL for interactive 3D plots
Animation Lets to make animated GIFs or SWF animations
Provides wrappers to half a dozen JavaScript plotting
rCharts
libraries using lattice syntax.
 HIGHLIGHTS
 Exploratory Data Analysis (EDA) shows how to use visualisation and
transformation to explore data in a systematic way.
 The main graphical packages are base, lattice and ggplot2.
 In R the pie chart is created using the pie() function.
 A 3D Pie Chart can be drawn using the package plotrix which uses the
function pie3D().
 The basic scatter plot in the base graphics system can be obtained by
using the plot() function.
 We use the arguments col and pch (values between 1 and 25) in the plot()
function to specify colour and plot pattern.
 The layout() function is used to control the layout of multiple plots in the
matrix.
 We use pairs() function to create matrices of scatter plots.

 The lattice graphics system has equivalent of plot() function and it is
xyplot().
 In the ggplot2 graphics, each plot is drawn with a call to the ggplot()
function.
 The ggplots can also be split into several panels like the lattice plots and
this is done using the function facet_wrap().
 The plot() function in R is used to create the line graph in base graphics
and the argument type = “l” is used to mention that it is a line.
 Line plots in the lattice graphics uses the xyplot() function and the
argument type = “l” is used to mention that it is a line.
 The ggplot2 scatter plot are created by adding the function geom_line() to
the ggplot() function.
 In R histograms in the base graphics are drawn using the function hist().
 The lattice histogram is drawn using the function histogram().
 The ggplot2 histograms are created by adding the function geom_
histogram() to the ggplot() function.
 In R base graphics the box plot is created using the boxplot() function.
 In the lattice graphics the box plot is drawn using the function bwplot().
 In the ggplot2 graphics the box plot is drawn by adding the function
geom_boxplot() to the function ggplot().
 In R base graphics the bar chart is created using the barplot() function.
 The lattice equivalent of the function barplot(), is the function barchart().
 In the ggplot2 graphics the bar chart is drawn by adding the function
geom_bar() to the function ggplot().
CHAPTER 5
Statistical Analysis Using R
 OBJECTIVES

 obtain the basic statistical measures like mean, median, mode, standard
deviation, variation etc., using R
 obtain the summary statistics of a given data
 understand and plot the normal distribution of data using R functions
 understand and plot the binomial distribution of data using R functions
 perform correlation analysis on the given data using R
 perform regression analysis on the given data using R
 perform ANOVA, ANCOVA on the given data using R
 perform chi-square and hypothesis testing on the given data using R
Statistical analysis in R is performed by using many in-built functions. Most of

these functions are part of the R base package. These functions take R vector as an
input along with the arguments and give the result. The other important R package
for statistical analysis is the stats package.
5.1. Basic Statistical Measures

Any dataset available in R or that is been imported into R for further analysis will
have both categorical data as well as numeric data. So, we can apply the statistical
functions available in R on the numeric data and understand the statistical
measures of the fields. The basic statistical measures are the minimum, maximum,
mean and median represented by the functions min(), max(), mean() and median()
respectively. Let us use the dataset named mtcars that is available in R by default to
understand these statistical measures.
> data(mtcars)
> colnames(mtcars)
[1] “mpg” “cyl” “disp” “hp” “drat” “wt” “qsec” “vs” “am” “gear” “carb”
> min(mtcars$cyl)
[1] 4
> max(mtcars$cyl)
[1] 8
> mean(mtcars$cyl)
[1] 6.1875
> median(mtcars$cyl)
[1] 6
All the above results can also be obtained by one function summary() and this
can also be applied on all the fields of the dataset at one shot. The range() function
gives the minimum and maximum values of a numeric field at one go.
> summary(mtcars$cyl)
4.000 4.000 6.000 6.188 8.000 8.000
> range(mtcars$cyl)
[1] 4 8
5.1.1. Mean
Mean is calculated by taking the sum of the values and dividing with the number of
values in a data series. The function mean() is used to calculate this in R. The basic
syntax for calculating mean in R is given below along with its parameters.
mean(x, trim = 0, na.rm = FALSE, ...)
x - numeric vector
143 Statistical Analysis Using R
trim - to drop some observations from both end of the sorted vector
na.rm - to remove the missing values from the input vector
> x <- c(45, 56, 78, 12, 3, -91, -45, 15, 1, 24)
> mean(x)
[1] 9.8
When trim parameter is supplied, the values in the vector get sorted and then
the required numbers of observations are dropped from calculating the mean.
When trim = 0.2, 2 values from each end will be dropped from the calculations to
find mean. In this case the sorted vector is (-91, -45, 1, 3, 12, 15, 24, 45, 56, 78) and
the values removed from the vector for calculating mean are (−91, −45) from left
and (56, 78) from right.
> mean(x, trim = 0.2)
[1] 16.66667
If there are missing values, then the mean() function returns NA. To drop the
missing values from the calculation use na.rm = TRUE, which means remove the
NA values.
> x <- c(45, 56, 78, 12, 3, -91, NA, -45, 15, 1, 24, NA)
> mean(x)
[1] NA
> mean(x, na.rm = TRUE)
[1] 9.8
5.1.2. Median
The middle most value in a data series is called the median. The median() function
is used in R to calculate this value. The basic syntax for calculating median in R is
given below along with its parameters.
median(x, na.rm = FALSE)
x - numeric vector
na.rm - to remove the missing values from the input vector
> x <- c(45, 56, 78, 12, 3, -91, -45, 15, 1, 24)
> median(x)
[1] 13.5
5.1.3. Mode
The mode is the value that has highest number of occurrences in a set of data.
Unlike mean and median, mode can have both numeric and character data. R does
not have a standard in-built function to calculate mode. So we create a user function
to calculate mode of a data set in R. This function takes the vector as input and
gives the mode value as output.
Mode <- function(x)
{
y <- unique(x)
y[which.max(tabulate(match(x, y)))]
}
> x <- c(1,2,3,4,5,5,5)

> Mode(x)
[1] 5
> ch <- c(“a”, “e”, “i”, “o”, “u”, “u”, “a”, “a”)
> Mode(ch)
[1] “a”
The function unique() returns a vector, data frame or array like x but with
duplicate elements/rows removed. The function match() returns a vector of
the positions of (first) matches of its first argument in its second. The function
tabulate() takes the integer-valued vector bin and counts the number of times each
integer occurs in it. The function which.max() determines the location, i.e., index
of the (first) maximum of a numeric (or logical) vector.
5.1.4. Standard Deviation and Variance
The functions to calculate the standard deviation, variance and the mean absolute
deviation are sd(), var() and mad() respectively.
> sd(mtcars$cyl)
[1] 1.785922
> var(mtcars$cyl)
[1] 3.189516
> mad(mtcars$cyl)
[1] 2.9652
5.1.5. Quartile Ranges
The quantile() function provides the quartiles of the numeric values. An alternative
function for quartiles is fivenum(). The IQR() function provides the inter quartile
range of the numeric fields.
> quantile(mtcars$cyl)
0% 25% 50% 75% 100%
4 4 6 8 8
> fivenum(mtcars$cyl)
[1] 4 4 6 8 8
> IQR(mtcars$cyl)
[1] 4
5.1.6. Other Statistical Functions
The function cor() and cov() are used to find the correlation and covariance between
two numeric fields respectively. In the below example the value shows that there is
negative correlation between the two numeric fields.
> cor(mtcars$mpg, mtcars$cyl)
[1] -0.852162
> cov(mtcars$mpg, mtcars$cyl)

[1] -9.172379
There are other statistics functions such as pmin(), pmax() [parallel equivalents
of min() and max() respectively], cummin() [cumulative minimum value], cummax()
[cumulative maximum value], cumsum() [cumulative sum] and cumprod()
[cumulative product].
> nrow(mtcars)
[1] 32
> mtcars$cyl
[1] 6 6 4 6 8 6 8 4 4 6 6 8 8 8 8 8 8 4 4 4 4 8 8 8 8 4 4 4 8 6 8 4
> pmin(mtcars$cyl)
[1] 6 6 4 6 8 6 8 4 4 6 6 8 8 8 8 8 8 4 4 4 4 8 8 8 8 4 4 4 8 6 8 4
> pmax(mtcars$cyl)
[1] 6 6 4 6 8 6 8 4 4 6 6 8 8 8 8 8 8 4 4 4 4 8 8 8 8 4 4 4 8 6 8 4
> cummin(mtcars$cyl)
[1] 6 6 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4
> cummax(mtcars$cyl)
[1] 6 6 6 6 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8
> cumsum(mtcars$cyl)
[1] 6 12 16 22 30 36 44 48 52 58 64 72 80 88 96 104 112 116 120 124
[21] 128 136 144 152 160 164 168 172 180 186 194 198
> cumprod(mtcars$cyl)
[1] 6.000000e+00 3.600000e+01 1.440000e+02 8.640000e+02 6.912000e+03
4.147200e+04
[7] 3.317760e+05 1.327104e+06 5.308416e+06 3.185050e+07 1.911030e+08
1.528824e+09
[13] 1.223059e+10 9.784472e+10 7.827578e+11 6.262062e+12 5.009650e+13
2.003860e+14
[19] 8.015440e+14 3.206176e+15 1.282470e+16 1.025976e+17 8.207810e+17

6.566248e+18
[25] 5.252999e+19 2.101199e+20 8.404798e+20 3.361919e+21 2.689535e+22
1.613721e+23
[31] 1.290977e+24 5.163908e+24
5.2. Summary Statistics

Thus the summary() function can be applied on the entire dataset to get all the
statistical values of all the numeric fields.
> summary(mtcars)
mpg cyl disp hp drat
Min. :10.40 Min. :4.000 Min. : 71.1 Min. : 52.0 Min. :2.760
1st Qu.:15.43 1st Qu.:4.000 1st Qu.:120.8 1st Qu.: 96.5 1st Qu.:3.080
Median :19.20 Median :6.000 Median :196.3 Median :123.0 Median :3.695
Mean :20.09 Mean :6.188 Mean :230.7 Mean :146.7 Mean :3.597
3rd Qu.:22.80 3rd Qu.:8.000 3rd Qu.:326.0 3rd Qu.:180.0 3rd Qu.:3.920
Max. :33.90 Max. :8.000 Max. :472.0 Max. :335.0 Max. :4.930
wt qsec vs am gear
Min. :1.513 Min. :14.50 Min. :0.0000 Min. :0.0000 Min. :3.000
1st Qu.:2.581 1st Qu.:16.89 1st Qu.:0.0000 1st Qu.:0.0000 1st Qu.:3.000
Median :3.325 Median :17.71 Median :0.0000 Median :0.0000 Median :4.000
Mean :3.217 Mean :17.85 Mean :0.4375 Mean :0.4062 Mean :3.688
3rd Qu.:3.610 3rd Qu.:18.90 3rd Qu.:1.0000 3rd Qu.:1.0000 3rd Qu.:4.000
Max. :5.424 Max. :22.90 Max. :1.0000 Max. :1.0000 Max. :5.000
carb
Min. :1.000
1st Qu.:2.000
Median :2.000
Mean :2.812
3rd Qu.:4.000
Max. :8.000
5.3. Normal Distribution

In a random collection of data from independent sources, it is generally observed
that the distribution of data is normal. Which means, on plotting a graph with the
value of the variable in the horizontal axis and the count of the values in the vertical
axis we get a bell shape curve. The centre of the curve represents the mean of the
data set. In the graph, half of values lie to the left of the mean and the other half lie
to the right of the graph. This is referred as normal distribution in statistics. R has
four in-built functions to generate normal distribution. They are described below.
dnorm(x, mean, sd)
pnorm(x, mean, sd)
qnorm(p, mean, sd)
rnorm(n, mean, sd)
x - vector of numbers
p - vector of probabilities
n - sample size
mean - mean (default value is 0)
sd - standard deviation (default value is 1)
5.3.1. dnorm()
For a given mean and standard deviation, this function gives the height of the
probability distribution. Below is an example in which the result of the dnorm()
function is plotted in a graph in Fig. 5.1.
> x <- seq(-5,5, by = 0.05)
> y <- dnorm(x, mean = 1.5, sd = 0.5)
> plot(x, y)
Figure 5.1 Plot of dnorm()
5.3.2. pnorm()
The pnorm() function returns the probability of a normally distributed random

number which is less than the value of a given number. The other name for this is
“Cumulative Distribution Function”. Below is an example in which the result of the
pnorm() function is plotted in a graph as in Fig. 5.2.
> x <- seq(-5,5, by = 0.05)
> y <- pnorm(x, mean = 1.5, sd = 1)
> plot(x, y)
Figure 5.2 Plot of pnorm()

5.3.3. qnorm()
The qnorm() function takes the probability value as input and returns a cumulative
value that matches the probability value. Below is an example in which the result of
the qnorm() function is plotted in a graph as in Fig. 5.3.
> x <- seq(0, 1, by = 0.02)
> y <- qnorm(x, mean = 2, sd = 1)
> plot(x, y)
Figure 5.3 Plot of qnorm()
5.3.4. rnorm()
This function is used to generate random numbers whose distribution is normal. It

takes the sample size as input and generates that many random numbers. We draw
a histogram to show the distribution of the generated numbers as in Fig. 5.4.
Figure 5.4 Histogram Using rnorm()

> x <- rnorm(80)
> hist(x, main = “Normal Distribution”)
5.4. Binomial Distribution

The probability of success of an event is found by the binomial distribution model and
this has only two possible outcomes in a series of experiments. For example, tossing of
a coin always gives a head or a tail. During the binomial distribution, the probability
of finding exactly 3 heads when tossing a coin for 10 times is estimated. R has four
in-built functions to generate binomial distribution. They are described below.
dbinom(x, size, prob)
pbinom(x, size, prob)
qbinom(p, size, prob)
rbinom(n, size, prob)
x - vector of numbers
p - vector of probabilities
n - sample size
size – number of trials
prob – probability of success of each trial
5.4.1. dbinom()
This function gives the probability density distribution at each point. Below is an
example in which the result of the dbinom() function is plotted in a graph as in Fig. 5.5.
> x <- seq(0, 25, by = 1)
> y <- dbinom(x,25,0.5)
> plot(x, y)
Figure 5.5 Plot Using dbinorm()
5.4.2. pbinom()
This function gives the cumulative probability of an event. It is a single value

representing the probability. The probability of getting 25 or less heads from a 50
tosses of a coin is given by the below code.
> x <- pbinom(25,50,0.5)
>x
[1] 0.5561376
5.4.3. qbinom()
The function qbinom() takes the probability value as input and returns a number
whose cumulative value matches the probability value. The below example finds how
many heads will have a probability of 0.5 will come out when a coin is tossed 50 times.
> x <- qbinom(0.5, 50, 1/2)
>x
[1] 25
5.4.4. rbinom()
The function rbinom() returns the required number of random values of the given
probability from a given sample. The below code is to find 5 random values from a
sample of 50 with a probability of 0.5.
> x <- rbinom(5,50,0.5)
>x
[1] 24 21 22 29 32
5.5. Correlation Analysis

To evaluate the relation between two or more variables, the correlation test is used.
Correlation coefficient in R can be computed using the functions cor() or cor.test().
The basic syntax for the correlation functions in R are as below.
cor(x, y, method)
cor.test(x, y, method)
x, y - numeric vectors with the same length

method - correlation method (“pearson” or “kendall” or “spearman”)
Consider the data set “mtcars” available in the R environment. Let us first find
the correlation between the horse power (“hp”) and the mileage per gallon (“mpg”)
of the cars and then between the horse power (“hp”) and the cylinder displacement
(“disp”) of the cars. From the test we find that the horse power (“hp”) and the
mileage per gallon (“mpg”) of the cars have negative correlation (-0.7761684) and
the horse power (“hp”) and the cylinder displacement (“disp”) of the cars have
positive correlation (0.7909486).
> cor(mtcars$hp, mtcars$mpg, method = “pearson”)
[1] -0.7761684
> cor.test(mtcars$hp, mtcars$mpg, method = “pearson”)

Pearson’s product-moment correlation
data: mtcars$hp and mtcars$mpg
t = -6.7424, df = 30, p-value = 1.788e-07
alternative hypothesis: true correlation is not equal to 0
95 percent confidence interval:
-0.8852686 -0.5860994
sample estimates:
cor
-0.7761684
> cor(mtcars$hp, mtcars$disp, method = “pearson”)

[1] 0.7909486
> cor.test(mtcars$hp, mtcars$disp, method = “pearson”)
Pearson’s product-moment correlation

data: mtcars$hp and mtcars$disp
t = 7.0801, df = 30, p-value = 7.143e-08
alternative hypothesis: true correlation is not equal to 0
0.6106794 0.8932775
sample estimates:
cor
0.7909486
The correlation results can also be viewed graphically as in Fig. 5.6. The corrplot()
function can be used to analyze the correlation between the various columns of a
dataset, say mtcars. After this, the correlation between individual columns can be
compared by plotting it in separate graphs as in Fig. 5.7 and Fig. 5.8.
> library(corrplot)
> M <- cor(mtcars)
> corrplot(M, method = “number”)
Figure 5.6 Cor Plot of mtcars Dataset

> plot(mtcars$hp, mtcars$mpg, xlab=”Horse Power of the Cars”,
ylab=”Mileage per Gallon of the Cars”, pch=21)
Figure 5.7 Scatter Plot of Negative Correlation

> plot(mtcars$hp, mtcars$disp, xlab=”Horse Power of the Cars”,

ylab=”Cylinder Displacement of the Cars”, pch=21)
Figure 5.8 Scatter Plot of Positive Correlation
It can be noted that the graph with negative correlation (Fig. 5.7) has the dots
from top left corner to bottom right corner and the graph with positive correlation
(Fig. 5.8) has the dots from the bottom left corner to the top right corner.
5.6. Regression Analysis
5.6.1. Linear Regression
Regression analysis is a widely used statistical tool. A relationship model is established

between the two variables used in the regression analysis. One of these variable
is called predictor variable whose value is gathered through experiments. The
other variable is called response variable whose value is derived from the predictor
variable. Linear Regression of the two variables are related through an equation.
The exponent of both the variables in this equation is 1. A linear relationship is
represented by a straight line when plotted as a graph. A non-linear relationship
where the exponent of any variable is not equal to 1 creates a curve.
The general mathematical equation for a linear regression is: y = ax + b, where

y is the response variable, x is the predictor variable and a and b are constants which
are called the coefficients.
A simple example of regression is predicting income of a person when his

expenditure is known. To do this we need to have the relationship between income
and expenditure of a person. First, carry out the experiment of gathering a sample
of observed values of income and corresponding expenditures. Then create a
relationship model using the lm() function in R. Find the coefficients from the
model created and create the mathematical equation using these values. Now, get a
summary of the relationship model to know the average error in prediction, which
is also called residuals. Finally, to predict the income of the new persons, use the
predict() function in R.
The function lm() creates the relationship model between the predictor and the
response variable. The basic syntax for lm() function in linear regression is as given
below.
lm(formula,data)
formula - relation between x and y

data - numeric vector
> x <- c(1510, 1740, 1380, 1860, 1280, 1360, 1790, 1630, 1520, 1310)
> y <- c(6300, 8100, 5600, 9100, 4700, 5700, 7600, 7200, 6200, 4800)
> model <- lm(y~x)
> model
Call:
lm(formula = y ~ x)
Coefficients:
(Intercept) x
-3845.509 6.746
> summary(model)
Call:
lm(formula = y ~ x)
Residuals:
Min 1Q Median 3Q Max

-630.02 -166.29 4.12 189.44 397.75
Coefficients:
Estimate Std. Error t value Pr(>|t|)
(Intercept) -3845.5087 804.9013 -4.778 0.00139 **
x 6.7461 0.5191 12.997 1.16e-06 ***
---
Signif. codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1
Residual standard error: 325.3 on 8 degrees of freedom

Multiple R-squared: 0.9548, Adjusted R-squared: 0.9491
F-statistic: 168.9 on 1 and 8 DF, p-value: 1.164e-06
The basic syntax for the function predict() in linear regression is as given below.
predict(object, newdata)
object - model created using lm() function

newdata - new numeric vector for predictor variable
> expense <- data.frame(x = 4700)
> income <- predict(model, expense)
> income
1
27861.18
The graphical representation of this linear regression is drawn by the below

code and Fig. 5.9.
> plot(y,x,col = “blue”,main = “Income & Expenditure Regression”,
abline(lm(x~y)),cex = 1.3,pch = 16,
xlab = “Income in Rs.”,ylab = “Expenditure in Rs.”)
Figure 5.9 Plot of Linear Regression
5.6.2. Multiple Regressions
Multiple regressions is an extension of linear regression into relationship between

more than two variables. In simple linear relation we have one predictor and one
response variable, but in multiple regressions we have more than one predictor
variable and one response variable.
The equation for multiple regressions consists of the below variables.

y = a + b1x1 + b2x2 +...bnxn
In this equation y is the response variable, a, b1, b2...bn are the coefficients and
x1, x2, ...xn are the predictor variables.
In R, the lm() function is used to create the regression model. The model
determines the value of the coefficients using the input data. Next we can predict
the value of the response variable for a given set of predictor variables using these
coefficients. The relationship model is built between the predictor variables and
the response variables. The basic syntax for lm() function in multiple regression is
as given below.
lm(y ~ x1+x2+x3..., data)
Consider the data set “mtcars” available in the R environment. This dataset
presents the data of different car models in terms of mileage per gallon (“mpg”),
cylinder displacement (“disp”), horse power (“hp”), weight of the car (“wt”) and
some more parameters. This model establishes the relationship between “mpg” as
a response variable with “disp”, “hp” and “wt” as predictor variables. We create a
subset of these variables from the mtcars data set for this purpose.
> model2 <- lm(mpg~disp+hp+wt, data = mtcars[, c(“mpg”, “disp”, “hp”, “wt”)])
> model2
Call:
lm(formula = mpg ~ disp + hp + wt, data = mtcars[, c(“mpg”, “disp”,
“hp”, “wt”)])
Coefficients:
(Intercept) disp hp wt
37.105505 -0.000937 -0.031157 -3.800891
> a <- coef(model2)[1]
>a
(Intercept)
37.10551
> b1 <- coef(model2)[2]
> b1
disp
-0.0009370091
> b2
hp
-0.03115655
> b3
wt
-3.800891
We create the mathematical equation below, from the above intercept and
coefficient values.
Y = a+b1*x1+b2*x2+b3*x3
Y = (37.10551)+(-0.0009370091)*x1+(-0.03115655)*x2+(-3.800891)*x3
We can use the regression equation created above to predict the mileage when
a new set of values for displacement, horse power and weight is provided. For a car
with disp = 160, hp = 110 and wt = 2.620 the predicted mileage is given by:
Y = (37.10551)+(-0.0009370091)*160+(-0.03115655)*110+(-3.800891)*2.620
= 23.57003
The above value can also be calculated using the function predict() for the given
new value.
> newdata <- data.frame(disp = 160, hp = 110, wt = 2.620)
> mileage <- predict(model2, newdata)
> mileage
1
23.57003
The graphical representation of this multiple regression is drawn by the below

code and in Fig. 5.10.
> plot(mtcars$mpg, mtcars$disp+mtcars$hp+mtcars$wt, col = “blue”,
main = “Mileage Per Gallons & Other Factors Multiple Regression”,
abline(lm(mtcars$disp+mtcars$hp+mtcars$wt~mtcars$mpg)),
cex = 1.5, pch = 15,
xlab = “Other Facors (mtcars$disp+mtcars$hp+mtcars$wt)”,
ylab = “Mileage Per Gallon”)
Figure 5.10 Plot of Multiple Regressions
5.6.3. Logistic Regression
The Logistic Regression is a regression model in which the response variable

(dependent variable) has categorical values such as True/False or 0/1. It actually
measures the probability of a binary response as the value of response variable based
on the mathematical equation relating it with the predictor variables. The general
mathematical equation for logistic regression is as given below.
y = 1 / ( 1 + e^ -(a + b1x1 + b2x2 + b3x3 + ... + bnxn))
The function used to create the logical regression model is the glm() function.
The basic syntax for glm() function in logistic regression is as given below.
glm(formula,data,family)
formula - relation between the variables x1, x2, ...xn and y

data - numeric vectors of data
family - takes the value “binomial” for logistic regression
The in-built data set “mtcars” describes different models of a car with their
various engine specifications. In “mtcars” data set, the column am describes the
transmission mode with a binary value. A logistic regression model is built between
the columns “am” and 3 other columns - hp, wt and cyl.
> model3 <- glm(am ~ cyl + hp + wt, data = mtcars[, c(“am”, “cyl”, “hp”, “wt”)],
family = binomial)
> model3
Call: glm(formula = am ~ cyl + hp + wt, family = binomial, data = mtcars[

c(“am”, “cyl”, “hp”, “wt”)])
Coefficients:
(Intercept) cyl hp wt
19.70288 0.48760 0.03259 -9.14947
Degrees of Freedom: 31 Total (i.e. Null); 28 Residual

Null Deviance: 43.23
Residual Deviance: 9.841 AIC: 17.84
> summary(model3)
Call:
glm(formula = am ~ cyl + hp + wt, family = binomial, data = mtcars[,
c(“am”, “cyl”, “hp”, “wt”)])
Deviance Residuals:
-2.17272 -0.14907 -0.01464 0.14116 1.27641
Coefficients:
Estimate Std. Error z value Pr(>|z|)
(Intercept) 19.70288 8.11637 2.428 0.0152 *
cyl 0.48760 1.07162 0.455 0.6491

hp 0.03259 0.01886 1.728 0.0840 .
wt -9.14947 4.15332 -2.203 0.0276 *
---
Signif. codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1
(Dispersion parameter for binomial family taken to be 1)

Null deviance: 43.2297 on 31 degrees of freedom
Residual deviance: 9.8415 on 28 degrees of freedom
AIC: 17.841
Number of Fisher Scoring iterations: 8
The p-value in the summary is greater than 0.05 for the variables “cyl” (0.6491)
and “hp” (0.0840). This value is considered to be insignificant in contributing to
the value of the variable “am”. Only weight “wt” (0.0276) impacts the “am” value
in this regression model.
5.6.4. Poisson Regression
Poisson Regression involves regression models in which the response variable is in

the form of counts and not fractional numbers. For example, the count of number
of births or number of wins in a football match series. Also the values of the response
variables follow a Poisson distribution. The Poisson Regression is represented by the
below equation.
log(y) = a + b1x1 + b2x2 + bnxn.....
The function used to create the Poisson regression model is the glm() function.
The basic syntax for glm() function in logistic regression is as given below.
glm(formula, data, family)
formula - relation between the variables x1, x2, ...xn and y

data - numeric vectors of data
family - takes the value “poisson” for poisson regression
The data set “warpbreaks” describes the effect of wool type and tension on the
number of warp breaks per loom. Let’s consider “breaks” as the response variable
which is a count of number of breaks. The wool “type” and “tension” are taken as
predictor variables. The model so built shows the below results.
> model4 <- glm(formula = breaks ~ wool + tension, data = warpbreaks,
family = poisson)
> summary(model4)
Call:
glm(formula = breaks ~ wool + tension, family = poisson, data = warpbreaks
Deviance Residuals:
-3.6871 -1.6503 -0.4269 1.1902 4.2616
Coefficients:
Estimate Std. Error z value Pr(>|z|)
(Intercept) 3.69196 0.04541 81.302 < 2e-16 ***
woolB -0.20599 0.05157 -3.994 6.49e-05 ***
tensionM -0.32132 0.06027 -5.332 9.73e-08 ***
tensionH -0.51849 0.06396 -8.107 5.21e-16 ***
---
Signif. codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1
(Dispersion parameter for poisson family taken to be 1)

Null deviance: 297.37 on 53 degrees of freedom
Residual deviance: 210.39 on 50 degrees of freedom
AIC: 493.06
Number of Fisher Scoring iterations: 4
In the summary we look for the p-value in the last column to be less than 0.05
to consider an impact of the predictor variable on the response variable. As seen the
wooltype B having tension type M and H have impact on the count of breaks.
p-value of woolB = 6.49e-05 = 0.0000649 < 0.05
p-value of tensionM = 9.73e-08 = 0.0000000973 < 0.05
p-value of tensionH = 5.21e-16 = 0.000000000000000521 < 0.05
5.6.5. Non-Linear Least Square Regression
When modelling real world data for regression analysis, we observe that it is rarely
the case that the equation of the model is a linear equation giving a linear graph. The
equation of the model of real world data involves mathematical functions of higher
degree. In such a scenario, the plot of the model gives a curve rather than a line. Both
linear and non-linear regression aims to adjust the values of the model’s parameters.
This is to find the line or curve that comes nearer to the data. On finding these values
we will be able to estimate the response variable with good accuracy.
In Least Square regression, a regression model is established. In this regression

model, the sum of the squares of the vertical distances of different points from the
regression curve is minimized. We generally start with a defined model and assume
some values for the coefficients. The nls() function of R is then apllied to get the
more accurate values and the confidence intervals. The basic syntax for creating a
nonlinear least square test in R is as below.
nls(formula, data, start)
formula - a nonlinear model formula including variables and parameters

data – a data frame
start - list or vector of starting estimates
We will consider a nonlinear model with assumption of initial values of its

coefficients. Next we will see what the confidence intervals of these assumed values
so that we can judge how well these values fit into the model. Consider the below
equation.
a = b1*x^2+b2
Let us assume the initial coefficients to be 1 and 3 and fit these values into nls()
function.
> x <- c(1.6, 2.1, 2, 2.23, 3.71, 3.25, 3.4, 3.86, 1.19, 2.21)
> y <- c(5.19, 7.43, 6.94, 8.11, 18.75, 14.88, 16.06, 19.12, 3.21, 7.58)
> plot(x, y)
> model <- nls(y ~ b1*x^2+b2, start = list(b1 = 1,b2 = 3))
> new <- data.frame(x = seq(min(x), max(x), len = 100))
> lines(new$x, predict(model, newdata = new))
> res1 <- sum(resid(model)^2)
> res1
[1] 1.081935
> res2 <- confint(model)
> res2
2.5% 97.5%
b1 1.137708 1.253135
b2 1.497364 2.496484
Figure 5.11 Plot of Non Linear Regressions

We can conclude that the value of b1 is more close to 1 (1.253135) while the
value of b2 is more close to 2 (2.496484) and not 3.
5.7. Analysis of Variance (ANOVA)

Analysis of Variance (ANOVA) is a statistical measure that is used for investigating
data by comparing the means of subsets of the data. The base case is the one-way
ANOVA. In one-way ANOVA the data is sub-divided into groups based on a single
classification factor. The one-way ANOVA is used to verify if the means of many
groups are equals. But this analysis may not be very useful for complex problems.
For example, it may be necessary to take into account two factors of variability to
determine if the averages between the groups depend on the group classification or
the second variable that is to consider. In this case the two-way analysis of variance
(two-way ANOVA) should be used.
Consider the dataset PlantGrowth available in R for performing one-way

ANOVA using R. This dataset has two columns, the control group / treatment and
the weight of the plant indicating its growth. We want to check if the hypothesis
that the control group / treatment has effect on the plant weight / plant growth. The
below code does the same.
> plant = lm(PlantGrowth$weight ~ PlantGrowth$group)
> anova(plant)
Analysis of Variance Table
Response: PlantGrowth$weight
Df Sum Sq Mean Sq F value Pr(>F)
PlantGrowth$group 2 3.7663 1.8832 4.8461 0.01591 *
Residuals 27 10.4921 0.3886
---
Signif. codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1
The result shows that the F-value is 4.8461 and the p-value is 0.01591 which is less
than 0.05 (5% level of significance). This shows that the null hypothesis is rejected,
that is the control group / treatment has effect on the plant growth / plant weight.
For two-way ANOVA, consider the below example of revenues collected for 5
years in each month. We want to see if the revenue depends on the Year and / or
Month or if they are independent of these two factors.
> revenue = c(15,18,22,23,24, 22,25,15,15,14, 18,22,15,19,21,
+ 23,15,14,17,18, 23,15,26,18,14, 12,15,11,10,8, 26,12,23,15,18,
+ 19,17,15,20,10, 15,14,18,19,20, 14,18,10,12,23, 14,22,19,17,11,
+ 21,23,11,18,14)
> months = gl(12,5)

> years = gl(5, 1, length(revenue))
> fit = aov(revenue ~ months + years)
> anova(fit)
Response: revenue
months 11 308.45 28.041 1.4998 0.1660
years 4 44.17 11.042 0.5906 0.6712
Residuals 44 822.63 18.696
The significance of the difference between months is: F = 1.4998. This value is
lower than the value tabulated and indeed p-value > 0.05. So we cannot reject the
null hypothesis: the means of revenue evaluated according to the months are not
proven to be not equal, hence we remain in our belief that the variable “months”
has no effect on revenue.
The significance of the difference between years is: F = 0.5906. This value is
lower than the value tabulated and indeed p-value > 0.05. So we fail to reject the
null hypothesis: the means of revenue evaluated according to the years are not found
to be un-equal, then the variable “years” has no effect on revenue.
5.8. Analysis of Covariance (ANCOVA)

We use Regression analysis to create models which describe the effect of variation
in predictor variables on the response variable. Sometimes, if we have a categorical
variable with values like Yes/No or Male/Female etc. the simple regression analysis
gives multiple results for each value of the categorical variable. In this case, we can
study the effect of categorical variable with the predictor variable. Such an analysis
is termed as Analysis of Covariance also called as ANCOVA.
ANCOVA is a type of ANOVA model that has a general linear model with a
continuous outcome variable and two or more predictor variables. Of these predictor
variables, at least one is continuous and at least one more is categorical. Analysis of
variance (ANOVA) is a collection of statistical models and their procedures which
are used to observe differences between the means of three or more variables in a
population based on the sample presented.
Consider the R built in data set “mtcars”. In this dataset the field “am”
represents the type of transmission and it takes the values 0 or 1. The miles per
gallon value, “mpg” of a car can also depend on it besides the value of horse power,
“hp”. The effect of the value of “am” on the regression between “mpg” and “hp” is
studied. It is done by using the aov() function followed by the anova() function to
compare the multiple regressions.
Consider the fields “mpg”, “hp” and “am” from the data set “mtcars”. The
variable “mpg” is the response variable, and the variable “hp” is chosen as the
predictor variable and “am” as the categorical variable. We create a regression model
taking “hp” as the predictor variable and “mpg” as the response variable taking into
account the interaction between “am” and “hp”.
The model with interaction between categorical variable and predictor variable
is given as below.
> res1 <- aov(mtcars$mpg ~ mtcars$hp * mtcars$am, data = mtcars)
> summary(res1)
mtcars$hp 1 678.4 678.4 77.391 1.50e-09 ***

mtcars$am 1 202.2 202.2 23.072 4.75e-05 ***
mtcars$hp:mtcars$am 1 0.0 0.0 0.001 0.981
Residuals 28 245.4 8.8
---
Signif. codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1
As the p-value in both the cases is less than 0.05, the result shows that both
horse power “hp” and transmission type “am” has significant effect on miles per
gallon “mpg”. But the interaction between these two variables is not significant as
the p-value is more than 0.05.
The model without interaction between categorical variable and predictor

variable is given as below.
> res2 <- aov(mtcars$mpg ~ mtcars$hp + mtcars$am, data = mtcars)
> summary(res2)
mtcars$hp 1 678.4 678.4 80.15 7.63e-10 ***
mtcars$am 1 202.2 202.2 23.89 3.46e-05 ***
Residuals 29 245.4 8.5
---
Signif. codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1
As the p-value in both cases is less than 0.05 in the above result, it shows that
both horse power and transmission type has significant effect on miles per gallon.
Now we can compare the two models to conclude if the interaction of the variables
is truly in-significant. For this we use the anova() function.
> finres <- anova(res1,res2)
> finres

Model 1: mtcars$mpg ~ mtcars$hp * mtcars$am

Model 2: mtcars$mpg ~ mtcars$hp + mtcars$am
Res.Df RSS Df Sum of Sq F Pr(>F)
1 28 245.43
2 29 245.44 -1 -0.0052515 6e-04 0.9806
As the p-value is (0.9806) greater than 0.05 we conclude that the interaction
between horse power “hp” and transmission type “am” is not significant. So the
mileage per gallon “mpg” will depend in a similar manner on the horse power of the
car in both auto and manual transmission mode.
5.9. Chi Square Test

Chi-Square test determines if two categorical variables have a significant correlation
between them. Both those variables should be from same population and they
should be categorical like − Yes/No, Male/Female, Red/Green etc. For example, we
can build a data set with observations on people’s ice-cream buying pattern and try
to correlate the gender of a person with the flavour of the ice-cream they prefer. If
a correlation is found we can plan for appropriate stock of flavours by knowing the
number of gender of people visiting.
The function chisq.test() is used for performing chi-Square test on the given
data. The R syntax for chi-square test is as below.
chisq.test(data)
data - data in form of a table
We will take the Cars93 data in the “MASS” library which represents the sales
of different models of car in the year 1993. The facor variables in this dataset can be
considered as categorical variables. In the below model the variables “AirBags” and
“Type” are considered. Here we aim to find out any significant correlation between
the types of car sold and the type of Air bags it has. If correlation is observed we can
estimate which types of cars can sell better with what types of air bags.
> library(MASS)
> cardata = table(Cars93$AirBags, Cars93$Type)
> chi <- chisq.test(cardata)
> chi
Pearson’s Chi-squared test
data: cardata
X-squared = 33.001, df = 10, p-value = 0.0002722
The result shows the p-value (0.0002723) of less than 0.05 which indicates a
strong correlation between the “AirBags” and “Type” of the cars sold.
5.10. Hypothesis Testing

The hypothesis is retained or rejected based on the p-value of the samples. This
decision is based on the statistical concept of hypothesis testing. A type I error is
the mishap of falsely rejecting a null hypothesis when the null hypothesis is true.
The significance level of the hypothesis testing is the probability of committing a
type I error. This is denoted by the Greek letter α (alpha). In hypothesis testing, a
type II error is due to a failure of rejecting an invalid null hypothesis. The probability
of avoiding a type II error is called the power of the hypothesis testing, and this is
denoted by the quantity 1 - β (1 – beta).
Hypothesis testing is of two types, one-tailed tests and two-tailed tests. A one-
tailed test is a test of a hypothesis which is used to measure the relationship of
variable in one direction which allows the hypothesis to be rejected. A two-tailed
test is used to test the statistical significance of the hypothesis for the dataset to
accept or reject the hypothesis.
5.10.1. Lower Tail Test

Suppose 60% of citizens voted in last election. 85 out of 148 people in a survey said
that they voted in current election. At 0.5 significance level, is it possible to reject
the null hypothesis that the proportion of voters in the population is above 60% this
year? We apply the prop.test() function to compute the p-value directly. The Yates
continuity correction is disabled.
> prop.test(85, 148, p=.6, alt=”less”, correct=FALSE)
1-sample proportions test without continuity correction
data: 85 out of 148, null probability 0.6

alternative hypothesis: true p is less than 0.6
0.0000000 0.6392527
sample estimates:
p
0.5743243
The p-value is 0.5743 > 0.05. Hence, at .05 significance level, we do not reject
the null hypothesis that the proportion of voters in the population is above 60%
this year.
5.10.2. Upper Tail Test
Let 12% of apples taken from an orchard last year were rotten. This year, 30 out of
214 apples turns out to be rotten. Is it possible to reject the null hypothesis at .05
level of significance, that the proportion of rotten apples in harvest stays below 12%
this year?
> prop.test(30, 214, p=.12, alt=”greater”, correct=FALSE)

alternative hypothesis: true p is greater than 0.12
0.1056274 1.0000000
sample estimates:
p
0.1401869
The p-value is 0.14018 > 0.05. Hence, at .05 significance level, we do not reject
the null hypothesis that the proportion of rotten apples in harvest stays below 12%
this year.
5.10.3. Two-Tailed Test
Let 12 heads are turned up out of 20 trials in a coin toss. At .05 significance level, is
it possible to reject the null hypothesis that the coin toss is fair?
> prop.test(12, 20, p=0.5, correct=FALSE)

alternative hypothesis: true p is not equal to 0.5
0.3865815 0.7811935
sample estimates:
p
0.6
The p-value is 0.6 > 0.05. Hence, at .05 significance level, we do not reject the
null hypothesis that the coin toss is fair.
 HIGHLIGHTS
 Most important functions to get the statistical measures are available in
the packages base and stats.
 The basic statistical measures are obtained by the functions min(), max(),
mean() and median().
 The basic statistical measures can also be obtained by one function
summary().
 R does not have a standard in-built function to calculate mode and hence
we create a user function to calculate mode.
 The function unique() returns a vector, data frame or array with duplicate
elements/rows removed.
 The function match() returns a vector of the positions of (first) matches
of its first argument in its second.
 The function tabulate() takes the integer-valued vector bin and counts
the number of times each integer occurs in it.
 The function which.max() determines the location, i.e., index of the (first)
maximum of a numeric (or logical) vector.
 The functions to calculate the standard deviation, variance and the mean
absolute deviation are sd(), var() and mad() respectively.
 The quantile() function provides the quartiles of the numeric values.
 The IQR() function provides the inter quartile range of the numeric fields.
 The function cor() and cov() are used to find the correlation and covariance
between two numeric fields respectively.
 R has the built in functions to generate normal distribution of data namely
dnorm(), pnorm(), qnorm() and rnorm().
 R has the built in functions to generate binomila distribution of data
namely dbinom(), pbinom(), qbinom() and rbinom().
 Correlation coefficient in R can be computed using the functions cor() or
cor.test().
 The corrplot() function plots the correlation between the various columns
of a dataset.
 The function lm() creates the relationship model between the predictor
and the response variable in linear / multiple regression analysis.
 Multiple regressions is an extension of linear regression where we have
more than one predictor variable and one response variable.
 The function used to create the logical / poisson regression model is the
glm() function.
 The function used for Analysis of Variance (ANOVA) / Analysis of
Covariance (ANCOVA) is anova() and aov().
 The function used for performing chi-Square test is chisq.test().
 In Hypothesis Testing, we apply the prop.test() function to compute the
p-value directly.
CHAPTER 6
Data Mining Using R
 OBJECTIVES

 know the main packages used for Data Mining in R
 know the packages and functions used for clustering in R
 apply k-means, k-medoids, hierarchical and density based clustering on
datasets
 know the packages and functions used for classification in R
 apply packages party, rpart and randomForest for classifying datasets
 perform association rule mining on datasets and visualize the results
 perform univariate / multivariate outlier detection using R
 perform outlier detection using LOF and clustering techniques in R
 perform principal component analysis using R
 perform feature selection using R
6.1. Packages for Data Mining

Data Mining comprises of techniques and algorithms, for determining interesting
patterns from large datasets. There are many algorithms that perform tasks such as
frequent pattern mining, clustering, and classification. It is a continuous challenge
faced by data mining analysts, to understand the importance of usage of these
algorithms. There are a large number of implementations available in R for the
various data mining algorithms currently present. Thus we can use the various
packages and function available in R for performing Clustering, Classification,
Association Rule Mining, Outlier Detection and Dimensionality Reduction. The

R packages that are related to data mining are stats, cluster, fpc, sna, e1071, cba,
biclust, clues, kohonen, rpart, party, randomForest, ada, caret, arules, eclat, arulesViz,
DMwR, dprep, Rlof, plyr, corrplot, RWeka, gausspred, optimsimplex, CCMtools,
FactoMineR and nnet.
6.2. Clustering Using R

Clustering is a process of grouping a set of objects into clusters in such a way that
objects in the same cluster are more similar to each other than to those in other
clusters. The various clustering techniques are Partitioning Method, Hierarchical
Method, Density-based Method, Grid-Based Method, Model-Based Method and
Constraint-based Method. R Package provides various packages and functions
that implement many clustering techniques such as K-Means, Hierarchical, Fuzzy
C-Means, PAM, SOM, CLARA, CLUES, etc. R-Package has many clustering
techniques implemented as the functions bundled in different library packages.
We can see the below table listing the clustering techniques available in R along
with their corresponding packages and functions.
6.2.1. Packages and Functions for Clustering
Clustering Algorithm Package Function

K-Means Clustering stats kmeans()
Partitioning Around Medoids cluster pam()
(PAM)
fpc pamk()
Hierarchical Clustering cluster agnes()
stats hclust()
Dissimilarity Matrix Calculation cluster daisy()
Density Based Clustering fpc dbscan()
K-Cores Clustering sna kcores()
179 Data Mining Using R

Fuzzy C-Means Clustering e1071 cmeans()
RockCluster cba rockCluster()
Biclust biclust biclust()
CLUES clues clues()
Self-Organizing Maps (SOM) kohonen som()
Proximus cba proximus()
CLARA cluster clara()
6.2.2. K-Means Clustering
K Means Clustering is an unsupervised learning algorithm. It tries to cluster data

based on their similarity. Unsupervised learning means that there is no outcome to
be predicted. The algorithm just tries to find patterns in the data. It is required to
specify the number of clusters needed in the data to be grouped. This algorithm
assigns randomly each observation to a cluster, and finds the cluster centroid. Then,
the algorithm repeats the below two steps: 1) Assign data to the cluster whose
centroid is closest. 2) For each cluster, calculates the new centroid. The above
two steps are repeated until the difference between the clusters cannot be further
reduced. The within cluster variation is calculated as the sum of the Euclidean
distance between the data and the cluster centroids.
R has the default dataset called iris and this dataset will be used in the example
below. The cluster number is set to 3 in the below clustering as the number of
distinct species in the iris dataset is 3 (“setosa”, “versicolor”, “virginica”). For the
purpose of initial manipulation let us copy the iris dataset into another dataset
called iris1. Then we remove the column “Species” from the dataset iris1 and then
apply kmeans clustering on it.
The result of the clustering is then compared with the “Species” column of the
dataset iris to see if similar objects are grouped together. The result of clustering
shows that the species “setosa” can be clustered separately and the other two species
have few overlapping of objects and hence can be clustered together. The functions
kmeans(), table(), plot() and point() are used below for getting and plotting the
results. It can be noted that the plots can be drawn with any two dimensions of the
species available at a particular time (eg. Sepal.Length Vs. Sepal.Width). Also, the
results of the kmeans clustering can vary from run to run due to the selection of
cluster centres.
> iris1 <- iris
> iris1$Species <- NULL
> km <- kmeans(iris1, 3)
> km
K-means clustering with 3 clusters of sizes 50, 38, 62
Cluster means:
1 5.006000 3.428000 1.462000 0.246000
2 6.850000 3.073684 5.742105 2.071053
3 5.901613 2.748387 4.393548 1.433871
Clustering vector:
[1] 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
[41] 1 1 1 1 1 1 1 1 1 1 3 3 2 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 2 3 3
[81] 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 2 3 2 2 2 2 3 2 2 2 2 2 2 3 3 2 2 2 2 3
[121] 2 3 2 3 2 2 3 3 2 2 2 2 2 3 2 2 2 2 3 2 2 2 3 2 2 2 3 2 2 3
Within cluster sum of squares by cluster:
[1] 15.15100 23.87947 39.82097
(between_SS / total_SS = 88.4 %)
Available components:
[1] “cluster” “centers” “totss” “withinss” “tot.withinss”
[6] “betweenss” “size” “iter” “ifault”
> table(iris$Species, km$cluster)
1 2 3
setosa 50 0 0
versicolor 0 2 48
virginica 0 36 14
> plot(iris1[c(“Sepal.Length”, “Sepal.Width”)], col = km$cluster)
> points(km$centers[, c(“Sepal.Length”, “Sepal.Width”)],
col = 1:3, pch = 8, cex = 2)
Figure 6.1 Plot of K-Means Clustering – Sepal.Length Vs. Sepal.Width

> plot(iris1[c(“Petal.Length”, “Petal.Width”)], col = km$cluster)
> points(km$centers[, c(“Petal.Length”, “Petal.Width”)], col = 1:3, pch = 8, cex = 2)
Figure 6.2 Plot of K-Means Clustering – Petal.Length Vs. Petal.Width
6.2.3. K-Medoids Clustering
The K-Medoids Clustering algorithm is a partitioning clustering algorithm related to

the K-Means algorithm. Both the K-Means and K-Medoids algorithms are partitional
(breaking the dataset up into groups) and both attempt to minimize the distance
between points labelled to be in a cluster and a point designated as the centre of
that cluster. In contrast to the K-Means algorithm, K-Medoids chooses data points as
centres and works with a generalization of the Manhattan Norm to define distance
between data points. This method clusters the data set of n objects into k clusters.
The silhouette is a useful tool for determining the number of clusters k.
R has the pam() and pamk() functions of the cluster package to do the k-medoids
clustering. The k-means and k-medoids clustering produces almost the same result
and the only difference is that in k-means the cluster is represented by the cluster
centre and in k-medoids the cluster is represented by the object close to the cluster
centre. But in the presence of outliers, k-medoids is more robust than k-means
clustering. Partitioning Around Medoids (PAM) is the classic algorithm applied for
k-medoids clustering. The PAM algorithm is not efficient in handling large datasets.
CLARA is an enhanced technique of PAM which performs better on large datasets.
For the functions pam() and clara() in the package cluster we need to specify the
number of clusters. But, for the function pamk() in the package fpc, we need not
specify the number of clusters. In this case the number of clusters is estimated
using the silhouette width.
> library(fpc)
> pmk <- pamk(iris1)
> pmk$nc
[1] 2
> table(pmk$pamobject$clustering, iris$Species)

1 50 1 0
2 0 49 50
> layout(matrix(c(1, 2), 1, 2))

> plot(pmk$pamobject)
> layout(matrix(1))
Figure 6.3 Plot of K-Medoids Clustering – Using pamk()
In the above left side chart of Fig 4.3, we can see that there are two clusters,
one for the species “setosa” and the other for the mixture of species “versicolor” and
“virginica”. The right side chart of Fig 4.3, shows the silhouette width which decides
the number of clusters (2 clusters in this case). The silhouette width is shown to be
between 0.81 and 0.62 (nearing 1), and this means that the observations are well
clustered. If the silhouette width is around 0 it means the observations lies between
the two clusters and it is less than 0, it means the observations are placed in the
wrong clusters.
Now, let us use the pam() function from the cluster package to cluster the iris
data and plot the results.
> library(cluster)
> pm <- pam(iris1, 3)
> table(pm$clustering, iris$Species)
1 50 0 0
2 0 48 14
3 0 2 36
> layout(matrix(c(1, 2), 1, 2))
> plot(pm)
> layout(matrix(1))
Figure 6.4 – Plot of K-Medoids Clustering – Using pam()

In the above left chart of Fig. 6.4, we can see three clusters. Cluster 1 with only
the “setosa” species and the cluster 2 with mostly “versicolor” species and few of
“virginica” species and in cluster 3 we have mostly “viginica” species and few of
“versicolor” species. In both the above graphs Fig. 6.3 and Fig. 6.4, the line between
the clusters shows the distance between the clusters. From the above we can say
that the choice of the clustering function used in R depends on the target problem
and the domain knowledge available.
6.2.4. Hierarchical Clustering
Hierarchical clustering is a method of cluster analysis which seeks to build a

hierarchy of clusters. Hierarchical clustering are of two types, Agglomerative and
Divisive. Agglomerative Hierarchical Clustering is a “bottom up” approach. In this
each observation starts with its own cluster, and pairs of clusters are merged into
one cluster and this moves up the hierarchy. Divisive Hierarchical Clustering is a
“top down” approach. In this all observations start in one cluster, and splits are
performed recursively as one cluster moves down the hierarchy. The results of
hierarchical clustering are usually presented in a dendrogram.
The hierarchical clustering in R can be done using the function hclust() in the
fpc package. As the hierarchical clustering when plotted will be very crowded if the
data is large, we create a sample of the iris data and do the clustering and plotting
using this sample data. The function rect.hclust() is used to draw rectangle that
covers each cluster. The cutree() function is used to draw the dendrogram as in the
Fig. 6.5.
> i <- sample(1:dim(iris)[1], 50)
> iris3 <- iris[i,]
> hc <- hclust(dist(iris3), method = “ave”)
> plot(hc, hang = -1, labels = iris$Species[i])
> rect.hclust(hc, k=3)
> grp <- cutree(hc, k=3)
Figure 6.5 Plot of Hierarchical Clustering – Using hclust()
The resultant graph Fig. 6.5, shows that the first cluster has just the species
“setosa”, the second cluster has the species “virginica” and the third cluster has a
mix of both the species “versicolor” and “virginica”.
6.2.5. Density Based Clustering
Density-based spatial clustering of applications with noise (DBSCAN) is a most

commonly used data clustering algorithm. In this clustering algorithm, given a set
of points in some space, it groups together points that are closely packed together,
marking as outliers points that lie alone in low-density regions.
The function dbscan() in the package fpc is used for the density based clustering.
The density based clustering is done to cluster the entire data into one cluster. There
are two main parameters in the function dbscan(). They are the eps and MinPts. The
parameter MinPts stands for the minimum points in a cluster and the parameter
eps defines the reachability distance. The density based clustering is sensitive to
noisy data. Standard values are given for these parameters eps and MinPts.
> library(fpc)
> iris4 <- iris
> db <- dbscan(iris4, eps = 0.42, MinPts = 5)
> table(db$cluster, iris$Species)

0 2 10 17
1 48 0 0
2 0 37 0
3 0 3 33
> plot(db, iris4)
In the result above we can see that there are three clusters, cluster 1, cluster 2
and cluster 3. The cluster 0 corresponds to the outliers in the data.
Figure 6.6 – Plot of Density Based Clustering – Using dbscan()
In the graph as in Fig. 6.6, the black circles represent the outliers. The clustering
results can also be plotted in the scatter plot like in k-means clustering. This can be
done using the function plot() as in Fig. 6.7, or the function plotcluster() as in Fig.
6.8, in the fpc package. The black circles in the Fig. 6.7 and the black zeros “0” in
the Fig. 6.8 shows the outliers.
> plot(db, iris4[c(1,2)])
> plotcluster(iris4, db$cluster)
Figure 6.7 Scatter Plot of Density Based Clustering – Using plot()
Figure 6.8 Scatter Plot of Density Based Clustering – Using plotcluster()
The clustering model can be used to model new data based on the similarity
between the new data and the clusters. We take a sample of 10 records from the iris
dataset and add some outliers to it and try to label the new dataset. The noises are
generated using the function runif().
> set.seed(435)
> i <- sample(1:dim(iris)[1], 10)
> irissamp <- iris[1]

> irissamp <- iris[i, ]
> irissamp$Species <- NULL
> irissamp <- irissamp + matrix(runif(10*4, min=0, max=0.3), nrow = 10, ncol = 4)
> plot(iris4[c(1,2)], col = db$cluster+1)
> points(irissamp[c(1,2)], pch=8, col=pred+1, cex = 2)
Figure 6.9 Scatter Plot of the Predicted Sample Clusters

> pred <- predict(db, iris4, irissamp)
> table(pred, iris$Species[i])
pred setosa versicolor virginica
0 0 0 1
1 3 0 0
2 0 2 0
3 0 2 2
Thus from the above results and the plot in Fig. 6.9, we can see that out of
the 10 new data, 7 (3 + 2 + 2) are assigned to the correct clusters and there is one
outlier data in it.
6.3. Classification
Classification is a data mining technique that assigns items in a sample to target
labelled classes. The goal of classification is to accurately predict the target class for
each case in the data. For example, a classification model could be used to identify
the age group of the people as children, youth, or old. The various classification
techniques are Decision Tree based Methods, Rule-based Methods, Memory
based reasoning, Neural Networks, Naïve Bayes and Bayesian Belief Networks and
Support Vector Machines. R Package provides various packages and functions that
implement many classification techniques such as SVM, kNN, Decision Trees,
Naive Bayes, etc. R-Package has many classification techniques implemented as the
functions bundled in different library packages. We can see the below table listing
the classification techniques available in R along with their corresponding packages
and functions.
6.3.1. Packages and Functions for Classification

SVM e1071 svm()
kNN RWeka IBk()
Decision Trees party rpart()
rpart ctree()
cforest()
randomForest randomForest()
Naïve Bayes e1071 naiveBayes()
Adaboost ada ada()
JRip caret train()
6.3.2. Decision Trees
Decision tree builds classification or regression models in the form of a tree

structure. The decision tree is incrementally developed with the dataset being
broken into smaller and smaller subsets. The final result is a tree with decision nodes
and leaf nodes. A decision node has two or more branches. Leaf node represents a
classification or decision. The topmost decision node in a tree is called the root
node and this corresponds to the best predictor. Decision trees can handle both
categorical and numerical data.
6.3.2.1. Package party
The function ctree() in the package party can be used to build the decision tree
for the given data. We consider the iris dataset available in R for our analysis. This
dataset has four attributes namely, the Sepal.Length, Sepal.Width, Petal.Length and
Petal.Width using which we can predict the Species of the flower. After applying the
function ctree() and getting the decision tree model, we can do the prediction using
the function predict() for the given new data, so that we can categorize it into which
Species the flowers belong to.
Before applying the decision tree function, the iris dataset is first split into
training and test subsets. For training we choose 80% of the data randomly and the
remaining 20% is used for testing. The seed for sampling is randomly set to a fixed
number as below for effective splitting of data. After creating the decision tree using
the training data, prediction is done on the test data. The results of the built tree
can be viewed as text result and as well as a decision tree plot as in Fig. 6.10.
> set.seed(1234)
> i <- sample(2, nrow(iris), replace=TRUE, prob=c(0.8, 0.2))
> train <- iris[i==1,]
> test <- iris[i==2,]
> form <- Species ~ Sepal.Length + Sepal.Width +
Petal.Length + Petal.Width
> dt <- ctree(form, data=train)
> table(predict(dt), train$Species)

setosa 42 0 0
versicolor 0 42 4
virginica 0 1 34
> plot(dt)
Figure 6.10 Decision Tree Plot – Using ctree()

> print(dt)
Conditional inference tree with 4 terminal nodes
Response: Species
Inputs: Sepal.Length, Sepal.Width, Petal.Length, Petal.Width
Number of observations: 123
1) Petal.Length <= 1.9; criterion = 1, statistic = 114.621
2)* weights = 42
1) Petal.Length > 1.9
3) Petal.Width <= 1.7; criterion = 1, statistic = 54.728
4) Petal.Length <= 4.8; criterion = 0.987, statistic = 8.717
5)* weights = 39
4) Petal.Length > 4.8
6)* weights = 7
3) Petal.Width > 1.7

7)* weights = 35
The above decision tree result can also be drawn as a simple decision tree as in
Fig. 6.11.
> plot(dt, type = “simple”)
Figure 6.11 Simple Decision Tree Plot – Using ctree()
In the first decision tree (Fig. 6.10) the number of training data under each
species is listed as bar graph, but in the second decision tree (Fig. 6.11) the same is
listed using variable y. For example, node 2 is labelled as “n = 42, y(1, 0, 0)”, which
means that it contains 42 training instances and all of them belong to the species
“setosa”.
Now, the predicted model will be tested with the test data to see if the instances
are correctly classified.
> pred <- predict(dt, newdata = test)
> table(pred, test$Species)
setosa 8 0 0
versicolor 0 7 1
virginica 0 0 11
6.3.2.2. Package rpart
The function rpart() in the package rpart can be used to build the decision tree for
the given data. We consider the bodyfat dataset available in the package TH.data of
R for our analysis. After applying the function rpart() and getting the decision tree
model, we can do the prediction using the function predict() for the given new data,
so that we can categorize it into which Species the flowers belong to.
Before applying the decision tree function, the dataset is first split into training
and test subsets. For training we choose 70% of the data randomly and the remaining
30% is used for testing. After creating the decision tree using the training data,
prediction is done on the test data. The decision tree is shown in the Fig. 6.12 and
the details of the split are listed below.
> library(TH.data)
> data(“bodyfat”, package = “TH.data”)
> set.seed(1234)
> i <- sample(2, nrow(bodyfat), replace=TRUE, prob=c(0.7, 0.3))
> train <- bodyfat[i==1,]
> test <- bodyfat[i==2,]
> form <- DEXfat ~ age + waistcirc + hipcirc +
elbowbreadth + kneebreadth
> dt2 <- rpart(form, data = train, control = rpart.control(minsplit = 10))
> plot(dt2)
> text(dt2, use.n=T, all = T, cex = 1)
Figure 6.12 – Classification Tree of bodyfat Dataset– Using rpart()

> str(bodyfat)
$ age : num 57 65 59 58 60 61 56 60 58 62 ...
$ DEXfat : num 41.7 43.3 35.4 22.8 36.4 ...
$ waistcirc : num 100 99.5 96 72 89.5 83.5 81 89 80 79 ...
$ hipcirc : num 112 116.5 108.5 96.5 100.5 ...
$ elbowbreadth: num 7.1 6.5 6.2 6.1 7.1 6.5 6.9 6.2 6.4 7 ...
$ kneebreadth : num 9.4 8.9 8.9 9.2 10 8.8 8.9 8.5 8.8 8.8 ...
$ anthro3a : num 4.42 4.63 4.12 4.03 4.24 3.55 4.14 4.04 3.91 3.66 ...
$ anthro3b : num 4.95 5.01 4.74 4.48 4.68 4.06 4.52 4.7 4.32 4.21 ...
$ anthro3c : num 4.5 4.48 4.6 3.91 4.15 3.64 4.31 4.47 3.47 3.6 ...
$ anthro4 : num 6.13 6.37 5.82 5.66 5.91 5.14 5.69 5.7 5.49 5.25 ...
> print(dt2)
n= 56
node), split, n, deviance, yval
* denotes terminal node
1) root 56 7265.0290000 30.94589
2) waistcirc< 88.4 31 960.5381000 22.55645
4) hipcirc< 96.25 14 222.2648000 18.41143

8) age< 60.5 9 66.8809600 16.19222 *
9) age>=60.5 5 31.2769200 22.40600 *
5) hipcirc>=96.25 17 299.6470000 25.97000
10) waistcirc< 77.75 6 30.7345500 22.32500 *
11) waistcirc>=77.75 11 145.7148000 27.95818
22) hipcirc< 99.5 3 0.2568667 23.74667 *
23) hipcirc>=99.5 8 72.2933500 29.53750 *
3) waistcirc>=88.4 25 1417.1140000 41.34880
6) waistcirc< 104.75 18 330.5792000 38.09111
12) hipcirc< 109.9 9 68.9996200 34.37556 *
13) hipcirc>=109.9 9 13.0832000 41.80667 *
7) waistcirc>=104.75 7 404.3004000 49.72571 *
> cp <- dt2$cptable[opt, “CP”]
> pred <- predict(dt2, newdata=test)
> xlim <- range(bodyfat$DEXfat)
> plot(pred ~ DEXfat, data=test, xlab=”Observed”,
ylab=”Predicted”, ylim=xlim, xlim=xlim)
> abline(a=0, b=1)
Figure 6.13 Graph of Predicted Vs. Observed Values

The predicted values are compared with the observed values and the graph in Fig.
6.13 shows that the modelling is good as most points lie close to the diagonal line.
Similarly, we can also apply the same rpart() function for the iris dataset as
before splitting the data into 80% training and 20% test data. The obtained
model (Fig. 6.14) can then be used for predicting the species of the test data. The
prediction shows that out of the 8 “setosa” species, all are correctly classified, out of
the 8 “versicolor” species, 7 are correctly classified and 1 is incorrectly classified as
“virginica” and out of the 11 “virginica” species, 10 are correctly classified and 1 is
incorrectly classified as “versicolor”.
> set.seed(1234)
> dt <- rpart(form, data = train, control = rpart.control(minsplit = 10))
> table(predict(dt), train$Species)
setosa 42 0 0
versicolor 0 42 4
virginica 0 1 34
> plot(dt)
> text(dt, use.n=TRUE, all=TRUE)
Figure 6.14 Classification Tree of Iris Dataset– Using rpart()

> print(dt)
n= 123
node), split, n, loss, yval, (yprob)
* denotes terminal node
1) root 123 80 versicolor (0.34146341 0.34959350 0.30894309)
2) Petal.Length< 2.45 42 0 setosa (1.00000000 0.00000000 0.00000000) *
3) Petal.Length>=2.45 81 38 versicolor
0.00000000 0.53086420 0.46913580)
6) Petal.Width< 1.75 46 4 versicolor
(0.00000000 0.91304348 0.08695652)
12) Petal.Length< 4.95 41 1 versicolor
(0.00000000 0.97560976 0.02439024) *
13) Petal.Length>=4.95 5 2 virginica
(0.00000000 0.40000000 0.60000000) *
7) Petal.Width>=1.75 35 1 virginica
(0.00000000 0.02857143 0.97142857) *
> pred <- predict(dt, newdata=test)

setosa 8 0 0
versicolor 0 7 1
virginica 0 1 10
6.3.2.3. Package randomForest
The function randomForest() in the package randomForest can be used to classify

the given data. We consider the same iris dataset available in R for our analysis.
The iris dataset as before is split into 80% training and 20% test data. The obtained
model (Fig. 6.15) can then be used for predicting the species of the test data.
> library(randomForest)
> set.seed(1234)
> rf <- randomForest(form, data=train, ntree=100, proximity=TRUE)
> plot(rf)
Figure 6.15 Random Forest of Iris Dataset

> table(predict(rf), train$Species)

setosa 42 0 0
versicolor 0 40 5
virginica 0 3 33
The importance of variables can be obtained with the functions importance()

and varImpPlot() as in the Fig. 6.16.
> importance(rf)
MeanDecreaseGini
Sepal.Length 7.328201
Sepal.Width 2.445821
Petal.Length 36.957590
Petal.Width 34.436098
> varImpPlot(rf)
Figure 6.16 Importance of Variables in Iris Dataset
Finally, the built random forest is tested on test data, and the result is checked
with the functions table() and margin(). The margin of a data point is the proportion
of the correct class minus maximum proportion of the other classes. Generally,
positive margin means correct classification (Fig. 6.17).
> pred <- predict(rf, newdata=test)

setosa 8 0 0
versicolor 0 7 0
virginica 0 0 12
> plot(margin(rf, test$Species))
Figure 6.17 Margin of Classification by Random Forest Using Iris Dataset
Hence, by comparing the predictions of the above three functions ctree(),

rpart() and randomForest(), we can see that the predictions of ctree() and rpart() are
almost same and randomForest() prediction is little different as the error rate (6.5%)
is high for the given data and sampling done
6.4. Association Rule Mining Using R

Association rules are rules presenting association or correlation between item sets.
An association rule is in the form of A => B, where A and B are two disjoint item
sets. A is called the LHS (left-hand side) and B is called the RHS (right-hand side)
of the rule. The three most widely-used measures for selecting interesting rules
are support, confidence and lift. Support is the percentage of cases in the data that
contains both A and B. Confidence is the percentage of cases containing A that also
contain B. Lift is the ratio of Confidence to the percentage of cases containing B.

The formulae to calculate them are as below.
support(A => B) = P(A U B)
confidence(A => B) = P(B|A) = P(A U B) / P(A)
lift(A => B) = confidence(A => B) / P(B) = P(A U B) / P(A) P(B)
Where P(A) is the percentage (or probability) of cases containing A and P(B) is
the percentage (or probability) of cases containing B.
The Titanic dataset in R is a 4-dimensional table with summarized information

on the fate of passengers on the Titanic according to social class, sex, age and survival.
We construct titanic.raw as raw data of the titanic dataset to make it suitable for
applying association rule mining. In this titanic.raw, each row represents a person.
> df <- as.data.frame(Titanic)
> titanic.raw <- NULL
> for(i in 1:4) { titanic.raw <- cbind(titanic.raw, rep(as.character(df[,i]), df$Freq)) }
> titanic.raw <- as.data.frame(titanic.raw)
> names(titanic.raw) <- names(df)[1:4]
> summary(titanic.raw)
Class Sex Age Survived
1st :325 Female: 470 Adult:2092 No :1490
2nd :285 Male :1731 Child: 109 Yes: 711
3rd :706
Crew:885
A classic algorithm for association rule mining is APRIORI. It is a level-wise,

breadth-first algorithm which counts transactions to find frequent item sets and
then derive association rules from them. An implementation of it is function
apriori() in package arules. Another algorithm for association rule mining is the
ECLAT algorithm, which finds frequent item sets with equivalence classes, depth-
first search and set intersection instead of counting. It is implemented using the
function eclat() in the package eclat.
In the below section, we do association rule mining using the function apriori().
This function has the default settings as supp=0.1, which is the minimum support
of rules, conf=0.8, which is the minimum confidence of rules and maxlen=10,
which is the maximum length of the rules.
> library(arules)
> rules.all <- apriori(titanic.raw)
> quality(rules.all) <- round(quality(rules.all), digits=3)
> rules.all
set of 27 rules
> inspect(rules.all)
lhs rhs support confidence lift
[1] {} => {Age=Adult} 0.950 0.950 1.000
[2] {Class=2nd} => {Age=Adult} 0.119 0.916 0.964
[3] {Class=1st} => {Age=Adult} 0.145 0.982 1.033
[4] {Sex=Female} => {Age=Adult} 0.193 0.904 0.951
[5] {Class=3rd} => {Age=Adult} 0.285 0.888 0.934
[6] {Survived=Yes}=> {Age=Adult} 0.297 0.920 0.968
[7] {Class=Crew} => {Sex=Male} 0.392 0.974 1.238
[8] {Class=Crew} => {Age=Adult} 0.402 1.000 1.052
[9] {Survived=No} => {Sex=Male} 0.620 0.915 1.164
[10] {Survived=No}=> {Age=Adult} 0.653 0.965 1.015
[11] {Sex=Male} => {Age=Adult} 0.757 0.963 1.013
[12] {Sex=Female,Survived=Yes}
=> {Age=Adult} 0.144 0.919 0.966
[13] {Class=3rd,Sex=Male}
=> {Survived=No} 0.192 0.827 1.222
[14] {Class=3rd,Survived=No}
=> {Age=Adult} 0.216 0.902 0.948
=> {Age=Adult} 0.210 0.906 0.953
[16] {Sex=Male,Survived=Yes}
=> {Age=Adult} 0.154 0.921 0.969
[17] {Class=Crew,Survived=No}
=> {Sex=Male} 0.304 0.996 1.266
[18] {Class=Crew,Survived=No}
=> {Age=Adult} 0.306 1.000 1.052
[19] {Class=Crew,Sex=Male}
=> {Age=Adult} 0.392 1.000 1.052
[20] {Class=Crew,Age=Adult}
=> {Sex=Male} 0.392 0.974 1.238
[21] {Sex=Male,Survived=No}
=> {Age=Adult} 0.604 0.974 1.025
[22] {Age=Adult,Survived=No}
=> {Sex=Male} 0.604 0.924 1.175
[23] {Class=3rd,Sex=Male,Survived=No}
=> {Age=Adult} 0.176 0.917 0.965
[24] {Class=3rd,Age=Adult,Survived=No}
=> {Sex=Male} 0.176 0.813 1.034
[25] {Class=3rd,Sex=Male,Age=Adult}
=> {Survived=No} 0.176 0.838 1.237
[26] {Class=Crew,Sex=Male,Survived=No}
=> {Age=Adult} 0.304 1.000 1.052
[27] {Class=Crew,Age=Adult,Survived=No}
=> {Sex=Male} 0.304 0.996 1.266
Many rules generated above are uninteresting. If we are interested in only rules
with rhs indicating survival, we set rhs=c(“Survived=No”, “Survived=Yes”) in the
appearance argument to make sure that only “Survived=No” and “Survived=Yes”

will appear in the rhs of rules. All other items can appear in the lhs, as set with the
default=”lhs”. In the above result rules.all, we can also see that the left-hand side
(lhs) of the first rule is empty. To exclude such rules, we set minlen to 2 in the
code below. The details of progress are suppressed with verbose=F. After applying
association rule mining, the resultant rules are sorted by lift so that the high-lift
rules appear first. In the below code, the minimum support is set to 0.005, so each
rule is supported at least by 12 (=ceiling(0.005 * 2201)) cases, which is acceptable
for a population of 2201.
> rules <- apriori(titanic.raw, control = list(verbose=F),
parameter = list(minlen=2, supp=0.005, conf=0.8),
appearance = list(rhs=c(“Survived=No”, “Survived=Yes”), default=”lhs”))
> quality(rules) <- round(quality(rules), digits=3)
> rules.sorted <- sort(rules, by=”lift”)
> inspect(rules.sorted)
[1] {Class=2nd,Age=Child}
=> {Survived=Yes} 0.011 1.000 3.096
[2] {Class=2nd,Sex=Female,Age=Child}
=> {Survived=Yes} 0.006 1.000 3.096
[3] {Class=1st,Sex=Female}
=> {Survived=Yes} 0.064 0.972 3.010
[4] {Class=1st,Sex=Female,Age=Adult}
=> {Survived=Yes} 0.064 0.972 3.010
[5] {Class=2nd,Sex=Female}
=> {Survived=Yes} 0.042 0.877 2.716
[6] {Class=Crew,Sex=Female}
=> {Survived=Yes} 0.009 0.870 2.692
[7] {Class=Crew,Sex=Female,Age=Adult}
=> {Survived=Yes} 0.009 0.870 2.692
[8] {Class=2nd,Sex=Female,Age=Adult}
=> {Survived=Yes} 0.036 0.860 2.663
[9] {Class=2nd,Sex=Male, Age=Adult}
=> {Survived=No} 0.070 0.917 1.354
[10] {Class=2nd,Sex=Male}
=> {Survived=No} 0.070 0.860 1.271
=> {Survived=No} 0.176 0.838 1.237
=> {Survived=No} 0.192 0.827 1.222
Some rules generated above provide little or no extra information when some
other rules are in the result. For example, the rule 2 provides no extra knowledge in
addition to rule 1, since rules 1 tells us that all 2nd-class children survived. A rule
is considered to be redundant when it is a super rule of another rule and it has the
same or a lower lift. Other redundant rules in the above result are rules 4, 7 and 8,
compared with the rules 3, 6 and 5 respectively. We prune redundant rules with
code below.
> subset.matrix <- is.subset(rules.sorted, rules.sorted)
> subset.matrix[lower.tri(subset.matrix, diag=T)] <- FALSE
> redundant <- colSums(subset.matrix, na.rm=T) >= 1
> which(redundant)
{Class=2nd,Sex=Female,Age=Child,Survived=Yes}
2
{Class=1st,Sex=Female,Age=Adult,Survived=Yes}
4
{Class=Crew,Sex=Female,Age=Adult,Survived=Yes}
7
{Class=2nd,Sex=Female,Age=Adult,Survived=Yes}
8
> rules.pruned <- rules.sorted[!redundant]
> inspect(rules.pruned)

[1] {Class=2nd,Age=Child}
=> {Survived=Yes} 0.011 1.000 3.096
[2] {Class=1st,Sex=Female}
=> {Survived=Yes} 0.064 0.972 3.010
[3] {Class=2nd,Sex=Female}
=> {Survived=Yes} 0.042 0.877 2.716
[4] {Class=Crew,Sex=Female}
=> {Survived=Yes} 0.009 0.870 2.692
[5] {Class=2nd,Sex=Male,Age=Adult}
=> {Survived=No} 0.070 0.917 1.354
[6] {Class=2nd,Sex=Male}
=> {Survived=No} 0.070 0.860 1.271
=> {Survived=No} 0.176 0.838 1.237
=> {Survived=No} 0.192 0.827 1.222
The above rules show that only 2nd class children have survived. But, this
cannot be the case as we have setup a higher support and confidence levels in
the previous case. To investigate the above issue, we run the code below to find
rules whose rhs is “Survived=Yes” and lhs contains “Class=1st”, “Class=2nd”,
“Class=3rd”, “Age=Child” and “Age=Adult” only, and which contains no other
items (default=”none”). We use lower thresholds for both support and confidence
than before to find all rules for children of different classes.
> rules <- apriori(titanic.raw, parameter = list(minlen=3, supp=0.002, conf=0.2),

appearance = list(rhs=c(“Survived=Yes”),
lhs=c(“Class=1st”, “Class=2nd”, “Class=3rd”, “Age=Child”, “Age=Adult”),
default=”none”), control = list(verbose=F))
> rules.sorted <- sort(rules, by=”confidence”)
> inspect(rules.sorted)
[1] {Class=2nd,Age=Child} => {Survived=Yes} 0.010904134 1.0000000 3.0956399
[2] {Class=1st,Age=Child} => {Survived=Yes} 0.002726034 1.0000000 3.0956399
[3] {Class=1st,Age=Adult} => {Survived=Yes} 0.089504771 0.6175549 1.9117275
[4] {Class=2nd,Age=Adult} =>{Survived=Yes} 0.042707860 0.3601533 1.1149048
[5] {Class=3rd,Age=Child} => {Survived=Yes} 0.012267151 0.3417722 1.0580035
[6] {Class=3rd,Age=Adult} => {Survived=Yes} 0.068605179 0.2408293 0.7455209
In the above result, the first two rules show that children of the 1st class are of
the same survival rate as children of the 2nd class and that all of them survived. The
rule of 1st-class children didn’t appear before, simply because of its support was
below the threshold specified. Rule 5 presents a sad fact that children of class 3 had
a low survival rate of 34%, which is comparable with that of 2nd-class adults and
much lower than 1st-class adults.
Now, let us see some ways of visualizing association rules such as Scatter Plot,
Grouped Matrix, Graph and Parallel Coordinates Plot as in Fig. 6.18, Fig. 6.19, Fig.
6.20 and Fig. 6.21 respectively.
> library(arulesViz)
> plot(rules.all)
> plot(rules.all, method=”grouped”)
Figure 6.18 Scatter Plot of 27 Rules from Titanic Dataset

> plot(rules.all, method=”graph”)
Figure 6.19 Grouped Matrix of 27 Rules from Titanic Dataset

Figure 6.20 Graph of 27 Rules from Titanic Dataset

> plot(rules.all, method=”paracoord”, control=list(reorder=TRUE))
Figure 6.21 Parallel Coordinates Plot of 27 Rules from Titanic Dataset
6.5. Outlier Detection Using R

An outlier is a data point that is distant from other data points. An outlier may be
caused due to difference in measuring or it may be due to error or oversight. There
are two types of outliers, namely, univariate and multivariate outliers. A univariate
outlier is an extreme data value of a particular variable. A multivariate outlier is
a combination of unusual data on minimum of two variables. These two types

of outliers can influence the outcome of the data analysis. Outliers exist for four
reasons. Incorrect data entry can cause data to contain extreme cases. Another
reason for outliers can be missing values in a dataset. Yet, another possibility is that
the data did not belong to the sample considered. And finally, the distribution of
the sample for specific variables may have a more extreme distribution than normal.
6.5.1. Univariate Outlier Detection
Univariate outlier detection is done with function boxplot.stats(), which returns

the statistics for producing boxplots. In the result returned by the function boxplot.
stats(), one component is out, which gives a list of outliers. It lists the data points
lying beyond the extremes of the whiskers. An argument of coef can be used to
control how far the whiskers extend out from the box of a boxplot. The result of the
same can be displayed as a box plot as in the Fig. 6.22.
> set.seed(3147)
> x <- rnorm(100)
> summary(x)
-3.3154 -0.4837 0.1867 0.1098 0.7120 2.6859
> boxplot.stats(x)$out
[1] -3.315391 2.685922 -3.055717 2.571203
> boxplot(x)
Figure 6.22 Box Plot of Univariate Outlier Detection
6.5.2. Multivariate Outlier Detection
The Univariate outlier detection can be used to find outliers in multivariate data
also. First, we create a data frame with the two independent variables and detect
their outliers separately. Then, we take multivariate outliers as those data which are
outliers for both variables. In the below code outliers are marked with “+” in red.
The result is displayed as a scatter plot as in Fig. 6.23.
> y <- rnorm(100)
> df <- data.frame(x, y)
> rm(x, y)
> head(df)
x y
1 -3.31539150 0.7619774
2 -0.04765067 -0.6404403
3 0.69720806 0.7645655
4 0.35979073 0.3131930
5 0.18644193 0.1709528
6 0.27493834 -0.8441813
> a <- which(df$x %in% boxplot.stats(df$x)$out)

> b <- which(df$y %in% boxplot.stats(df$y)$out)
>a
[1] 1 33 64 74
>b
[1] 24 25 49 64 74
> outlierlist1 <- intersect(a,b)
> outlierlist1
[1] 64 74
> plot(df)
> points(df[outlierlist1,], col=”red”, pch=”+”, cex=2.5)
Figure 6.23 Scatter Plot of Multivariate Outlier Detection Using intersect()
Similarly, we can take multivariate outliers as those data which are outliers in
either of the variables (x or y). This is shown as scatter plot in the Fig. 6.24.
> outlierlist2 <- union(a,b)
> outlierlist2
[1] 1 33 64 74 24 25 49
> plot(df)
> points(df[outlier.list2,], col=”blue”, pch=”x”, cex=2)
Figure 6.24 Scatter Plot of Multivariate Outlier Detection Using union()
When there are three or more variables in an application, a final list of outliers
might be produced with majority voting of outliers detected from individual
variables. Domain knowledge should be involved when choosing the optimal way to
group the items in real-world applications.
6.5.3. Outlier Detection Using LOF
LOF (Local Outlier Factor) is an algorithm for identifying density-based local

outliers. With LOF, the local density of a point is compared with that of its
neighbours. If the local density of this point is significantly lower than its neighbour
(with an LOF value greater than one), the point is in a sparser region than its
neighbours, which suggests it be an outlier. LOF works only on numeric data.
Function lofactor() calculates local outlier factors using the LOF algorithm, and
it is available in packages DMwR and dprep. An example of outlier detection with
LOF is given below, where k is the number of neighbours used for calculating local
outlier factors. A density plot of the outlier scores is also shown in the Fig. 6.25.
> library(DMwR)
> iris1 <- iris[,1:4]

> outscores <- lofactor(iris1, k=5)
> plot(density(outscores))
> outliers <- order(outscores, decreasing=T)[1:5]

> print(outliers)
[1] 42 107 23 110 63
> print(iris1[outliers,])
42 4.5 2.3 1.3 0.3
107 4.9 2.5 4.5 1.7
23 4.6 3.6 1.0 0.2
110 7.2 3.6 6.1 2.5
63 6.0 2.2 4.0 1.0
Figure 6.25 Density Plot of Outliers Detection Using lofactor()
We show outliers with a biplot of the first two principal components in the
Fig. 6.26. In the below code, prcomp() performs a principal component analysis,
and biplot() plots the data with its first two principal components. In the below
graph, Fig. 6.26, x-axis and y-axis are respectively the first and second principal
components, the arrows show the original columns (variables), and the five outliers
are labelled with their row numbers.
> n <- nrow(iris1)
> labels <- 1:n
> labels[-outliers] <- “.”

> biplot(prcomp(iris1), cex=.8, xlabs=labels)
We can also show outliers with a pairs plot as below, where outliers are labelled
with “+” in red.
Figure 6.26 Biplot of First Two Principal Components

> pch <- rep(“*”, n)
> pch[outliers] <- “+”
> col <- rep(“blue”, n)
> col[outliers] <- “red”
> pairs(iris1, pch=pch, col=col, cex = 3)
The outliers can also be displayed using the pairs() function as in the Fig. 6.27
in which the outliers are marked as “+” symbol displayed in red colour. Package Rlof
provides the function lof(), a parallel implementation of the LOF algorithm.
Figure 6.27 Outliers Displayed Using pairs()
6.5.4. Outlier Detection Using Clustering
One way to detect outliers is clustering. By grouping data into clusters, those
data not assigned to any clusters are taken as outliers. For example, with density-
based clustering such as DBSCAN, objects are grouped into one cluster if they
are connected to one another by densely populated area. Therefore, objects not
assigned to any clusters are isolated from other objects and are taken as outliers.
The function dbscan() in the package fpc is used for the density based clustering.
There are two main parameters in the function dbscan(). They are the eps and
MinPts. The parameter MinPts stands for the minimum points in a cluster and the
parameter eps defines the reachability distance. Standard values are given for these
parameters.
> library(fpc)
> iris1 <- iris
> db <- dbscan(iris1, eps = 0.42, MinPts = 5)
> table(db$cluster, iris$Species)

0 2 10 17
1 48 0 0
2 0 37 0
3 0 3 33
> plot(db, iris1)
In the result above we can see that there are three clusters, cluster 1, cluster 2
and cluster 3. The cluster 0 corresponds to the outliers in the data (marked by black
circles in the plot of Fig. 6.28).
Figure 6.28 Plot of Outliers Detection Using Density Based Clustering
The clustering results can also be plotted in the scatter plot as in Fig. 6.29. This
can be done using the function plot() in the fpc package. Here also the black circles
denote the outliers.
> plot(db, iris1[c(1,2)])
Figure 6.29 Scatter Plot of Outliers Detection Using Density Based Clustering
We can also detect outliers with the k-means algorithm. With k-means, the
data are partitioned into k groups by assigning them to the closest cluster centres.
After that, we can calculate the distance (or dissimilarity) between each object and
its cluster centre, and pick those with largest distances as outliers. An example of
outlier detection with k-means from the iris data is given below. In the graph of Fig.
6.30, the cluster centres are labelled with “*” and outliers with “+”.
> iris2 <- iris[,1:4]
> km <- kmeans(iris2, centers=3)
> centers <- km$centers[km$cluster, ]
> dist <- sqrt(rowSums((iris2 - centers)^2))
> outliers <- order(dist, decreasing=T)[1:5]
> print(outliers)
[1] 99 58 94 61 119
> print(iris2[outliers,])

99 5.1 2.5 3.0 1.1
58 4.9 2.4 3.3 1.0
94 5.0 2.3 3.3 1.0

61 5.0 2.0 3.5 1.0
119 7.7 2.6 6.9 2.3
> plot(iris2[,c(“Sepal.Length”, “Sepal.Width”)], pch=”o”, col=km$cluster, cex=0.3)
> points(km$centers[,c(“Sepal.Length”, “Sepal.Width”)], col=1:3, pch=8, cex=1.5)
> points(iris2[outliers, c(“Sepal.Length”, “Sepal.Width”)], pch=”+”, col=4, cex=1.5)
Figure 6.30 Scatter Plot of Outliers Detection Using K-Means Clustering
6.6. Dimensionality Reduction Using R

Dimensionality reduction is the process of reducing the number of irrelevant
variables in a given dataset for future predictive modelling. It is a very important
step of predictive modelling. If the idea is to improve accuracy of the model, it’s the
step where one needs to invest more time. Right variables produce more accurate
models. Hence, it is required to choose a strategy for variable selection for efficient
prediction from the given data. Some of these strategies are Principal Component
Analysis, Singular Value Decomposition and Feature Selection.
6.6.1. Principal Component Analysis
Sparse data is a commonly faced problems in data analytics. At many times, we

face a situation where we have a large set of features and fewer data points, or we
have data with very high feature vectors. In such scenarios, fitting a model to the
dataset results in lower predictive power of the model. This scenario is often termed
as the curse of dimensionality. Principal Component Analysis (PCA) is a popular
dimensionality reduction technique. PCA is used in many applications that deals
with high dimension of data.
Consider the data, we want to work with, is in the form of a matrix A of m x n

dimensions, where Ai,j represents the value of the ith observation of the jth variable.
Thus the n members of the matrix can be identified with the m rows, each variable
corresponding to n-dimensional vectors. If n is very large it is often desirable to
reduce the number of variables to a smaller number of variables, say k variables,
where k < n, while losing as little information as possible. PCA is a linear orthogonal
transformation that transforms the data to a new coordinate system. The projection
of the data lies on the first coordinate called as the first principal component.
The second transformation happens on the second coordinate called the second
principal component, and so on. The variables that have less variance are discarded
by the PCA technique.
PCA uses orthogonal projection of highly correlated variables to a set of values

of linearly uncorrelated variables called Principal Components. The number of
reduced principal components / features is less than or equal to the number of
original variables / features. The transformation is done in such a way that the first
principal component has the largest variance. By considering the highly correlated
features, the data varies as much as possible. Every next component has the highest
variance. This uses the features that are less correlated with the first principal
component and that are orthogonal to the preceding component.
We can implement PCA in R by using the crimtab dataset available in R. This

dataset contains the data of 3000 male criminals over 20 years old undergoing their
sentences in the chief prisons of England and Wales. The row names correspond to
midpoints of intervals of finger lengths whereas the column names correspond to
heights of 3000 criminals.
> str(crimtab)
‘table’ int [1:42, 1:22] 0 0 0 0 0 0 1 0 0 0 ...
- attr(*, “dimnames”)=List of 2
..$ : chr [1:42] “9.4” “9.5” “9.6” “9.7” ...

..$ : chr [1:22] “142.24” “144.78” “147.32” “149.86” ...
> dim(crimtab)
[1] 42 22
> sum(crimtab)
[1] 3000
> nrow(crimtab)
[1] 42
> colnames(crimtab)
[1] “142.24” “144.78” “147.32” “149.86” “152.4” “154.94” “157.48” “160.02”
“162.56” “165.1”
[11] “167.64” “170.18” “172.72” “175.26” “177.8” “180.34” “182.88” “185.42”
“187.96” “190.5”
[21] “193.04” “195.58”
> rownames(crimtab)
[1] “9.4” “9.5” “9.6” “9.7” “9.8” “9.9” “10” “10.1” “10.2” “10.3” “10.4” “10.5”
“10.6”
[14] “10.7” “10.8” “10.9” “11” “11.1” “11.2” “11.3” “11.4” “11.5” “11.6” “11.7”
“11.8” “11.9”
[27] “12” “12.1” “12.2” “12.3” “12.4” “12.5” “12.6” “12.7” “12.8” “12.9” “13”
“13.1” “13.2”
[40] “13.3” “13.4” “13.5”
Let us use the function apply() to the crimtab dataset row wise to calculate the
variance to see how each variable is varying. The function apply() returns a vector
or array or list of values obtained by applying a function to margins of an array or
matrix.
> apply(crimtab,2,var)
142.24 144.78 147.32 149.86 152.4
0.02380952 0.02380952 0.17421603 0.88792102 2.56445993
154.94 157.48 160.02 162.56 165.1

11.19860627 38.04471545 107.76596980 205.79616725 270.58536585
167.64 170.18 172.72 175.26 177.8
238.42973287 187.60569106 79.57491289 42.85540070 12.51161440
180.34 182.88 185.42 187.96 190.5
3.74680604 0.68583043 0.19105691 0.08826945 0.00000000
193.04 195.58
0.00000000 0.02380952
We can see that the column “165.1” contains maximum variance “270.58536585”.
Let us now apply PCA using the function prcomp() on the data set crimtab.
> pca =prcomp(crimtab)
> pca
Standard deviations (1, .., p=22):
[1] 30.07962021 14.61901911 5.45438277 4.65250574 3.21408168 2.77322835
2.30250353
[8] 1.92678188 1.40986049 1.24320894 1.02967875 0.72502776 0.50683548
0.47841947
[15] 0.29167315 0.26636232 0.22462458 0.12793888 0.12483426 0.06548509
0.00000000
[22] 0.00000000
Rotation (n x k) = (22 x 22):

PC1 PC2 PC3 PC4 PC5 PC6
142.24 0.0006121303 -9.156788e-05 0.0019577283 -0.002856530 0.0009242447 - 0.002241331
144.78 0.0005879745 -4.169480e-04 -0.0012002251 -0.001636190 0.0037931852 -0.00111885
...
...
From the above code, the resultant components of PCA object are the standard
deviations and Rotation. From the standard deviations we can observe that the 1st
PCA explained most of the variation, followed by other PCAs’. The proportion
of each variable along each principal component is given by the Rotation of the
principal component. Let’s plot all the principal components and see how the
variance is accounted with each component in Fig. 6.31.
> par(mar = rep(2, 4))
> plot(PCA)
Figure 6.31 Plot of Variances of the Principal Components
Clearly the first principal component accounts for maximum information. The
results of PCA can be represented as a biplot graph. Biplot is used to show the
proportions of each variable along the first two principal components as in Fig. 6.32.
The first two lines of the below code changes the direction of the biplot. If we do
not include the first two lines the plot will be mirror image of the below graph.
> PCA$rotation=-PCA$rotation
> PCA$x=-PCA$x
> biplot (PCA, scale = 0.2)
The below Fig. 6.32, is known as a biplot. In this, we can see the two principal
components (PC1 and PC2) of the crimtab dataset ploted in the graph. The arrows
represent the loading vectors, and this specifies how the feature space varies along
the principal component vectors. From the below plot, we can see that the first
principal component vector, PC1, more or less places equal weight on three features:
165.1, 167.64, and 170.18. This means that these three features are more correlated
with each other. In the second principal component, PC2 places more weight on
160.02, 162.56 than the other 3 features.
Figure 6.32 Biplot of the First Two Principal Components
6.6.2. Feature Selection
“Feature Selection” or “Variable Selection “ is a process of reducing the number of

variables that are irrelevant.
Step I: Remove Redundant Variables
The following 3 simple analysis helps to remove redundant variables:
1. Remove variables having high percentage of missing values
2. Remove Zero and Near Zero-Variance Predictors
3. Remove highly correlated variables (greater than 0.7)
The absolute values of pair-wise correlations are considered. Then, we remove

the variable with the largest mean absolute correlation looking at the mean absolute
correlation of each variable, if the two variables have high correlation. The below
code in R does the above steps.
> library(caret)
> library(corrplot)
> library(plyr)
> dat <- read.csv(“Sample.csv”)
> set.seed(227)
# Remove variables having high percentage of missing values
> dat1 <- dat[, colMeans(is.na(dat)) <= .5]
> dim(dat1)
[1] 19622 93
> dim(dat)
[1] 19622 160
> nzv <- nearZeroVar(dat1)
# Remove Zero and Near Zero-Variance Predictors
> dat2 <- dat1[, -nzv]
> dim(dat2)
[1] 19622 59
> numericData <- dat2[sapply(dat2, is.numeric)]
> summary(descrCor[upper.tri(descrCor)])
-0.992008 -0.101969 0.001729 0.001405 0.084718 0.980924
> corrplot(descrCor, order = “FPC”, method = “color”, type = “lower”,
tl.cex = 0.7, tl.col = rgb(0, 0, 0))
Figure 6.33 – Correlation Plot of the Features

> highlyCorrelated <- findCorrelation(descrCor, cutoff=0.7)
> print(highlyCorrelated)
[1] 14 5 13 26 8 7 40 12 15 41 39 42 43 25 38 27 29 16 52 50 49 37 22
> highlyCorCol <- colnames(numericData)[highlyCorrelated]
# Remove highly correlated variables (greater than 0.7)

> dat3 <- dat2[, -which(colnames(dat2) %in% highlyCorCol)]
> dim(dat3)
[1] 19622 36
Step II: Feature Selection with Random Forest
Random Forest is one of the most popular algorithm in data science. The
best part of this algorithm is there are no assumptions attached to it.
The pre-processing work is very less as compared to other techniques. It
overcomes the problem of over fitting that decision tree has. It provides a
list of predictor (independent) variables which are important in predicting
the target (dependent) variable. It contains two measures of variable
importance. The first one - Gini Gain produced by the variable, averaged
over all trees. The second one - Permutation Importance i.e. mean decrease
in classification accuracy after permuting the variable, averaged over all
trees. Sort the permutation importance score on descending order and
select the TOP k variables. The below code is for feature selection with
Random Forest.
> rf <-randomForest(classe~.,data=dat3, importance=TRUE,ntree=100)
# Finding importance of variables
> imp = importance(rf, type=1)
> imp <- data.frame(predictors=rownames(imp),imp)
# Sorting the variables in descending order of their MeanDecreaseAccuracy
> imp.sort <- arrange(imp,desc(MeanDecreaseAccuracy))
> imp.sort$predictors <-
factor(imp.sort$predictors,levels=imp.sort$predictors)
> imp.20<- imp.sort[1:20,]
# Printing top 20 variables with high MeanDecreaseAccuracy
> print(imp.20)
predictors MeanDecreaseAccuracy
1 X 36.878224
2 raw_timestamp_part_1 19.939217
3 cvtd_timestamp 19.936367
4 pitch_belt 14.474235
5 roll_dumbbell 12.502391
6 gyros_belt_z 12.429689
7 num_window 11.491461
8 total_accel_dumbbell 11.193014
9 gyros_arm_y 10.509349
10 magnet_forearm_z 10.353922
11 gyros_dumbbell_x 10.245442
12 magnet_belt_z 10.078787
13 pitch_forearm 10.069103
14 roll_arm 10.049374
15 yaw_arm 9.959173
16 gyros_dumbbell_y 9.770771
17 gyros_belt_x 9.602383
18 magnet_forearm_y 9.407758
19 user_name 9.304626
20 gyros_forearm_x 8.954952
> varImpPlot(rf, type=1)
# Retaining only top 20 variables with high MeanDecreaseAccuracy in dat4
> dat4 = cbind(classe = dat3$classe, dat3[,c(imp.20$predictors)])
To fix the number of features to be selected based on the ranking, a threshold

(say ‘n’) is fixed and based on this the top n (20) features are selected for further
analysis.
Figure 6.34 Correlation Plot of the Features

 HIGHLIGHTS
 The R packages that are related to data mining are stats, cluster, fpc,
sna, e1071, cba, biclust, clues, kohonen, rpart, party, randomForest, ada,
caret, arules, eclat, arulesViz, DMwR, dprep, Rlof, plyr, corrplot, RWeka,
gausspred, optimsimplex, CCMtools, FactoMineR and nnet.
 The R packages and function for clustering are stats - kmeans(), cluster
- pam(), fpc - pamk(), cluster - agnes(), stats - hclust(), cluster - daisy(),
fpc - dbscan(), sna - kcores(), e1071 - cmeans(), cba - rockCluster(), biclust
- biclust(), clues - clues(), kohonen - som(), cba - proximus() and cluster -
clara().
 The functions kmeans(), table(), plot() and point() are used below for
getting and plotting the results of K-Means clustering.
 R has the pam() and pamk() functions of the cluster package to do the
k-medoids clustering.
 The silhouette width decides the number of clusters in k-medoids
clustering.
 The hierarchical clustering in R can be done using the function hclust() in
the fpc package.
 The function dbscan() in the package fpc is used for the density based
clustering.
 The R packages and function for clustering are e1071 - svm(), RWeka
- IBk(), party - rpart(), rpart - ctree(), rpart - cforest(), randomForest -
randomForest(), e1071 - naiveBayes(), ada - ada(), caret - train().
 The function ctree() in the package party and the function rpart() in the
package rpart can be used to build the decision tree for the given data.
 We can do the prediction using the function predict() for the given new
data.
 The function randomForest() in the package randomForest can be used to
classify the given data.
 The importance of variables in a dataset can be obtained with the functions
importance() and varImpPlot().
 The built random forest is tested on test data, and the result is checked
with the functions table() and margin().
 An implementation of Association Rule Mining using APRIORI algorithm

is the function apriori() in package arules.
 An implementation of Association Rule Mining using ECLAT algorithm
is the function eclat() in package eclat.
 The function inspect() can be used to view the rules generated by the
function apriori().
 The package arulesViz is used to visualize the association rules as graphs
and plots.
 Univariate outlier detection is done with function boxplot.stats().
 Function lofactor() available in the packages DMwR and dprep calculates
local outlier factors using the LOF algorithm.
 The function prcomp() performs a principal component analysis, and
biplot() plots the data with its first two principal components.
 Package Rlof provides the function lof(), a parallel implementation of the
LOF algorithm.
CHAPTER 7
Case Studies
 OBJECTIVES

 understand how text mining is done using R
 understand how modelling can be done using R for credit risk analysis
 understand how EDA can be conducted on the two data sets, the crime data
and the base ball data
 understand how social network analysis can be done using R
7.1. Text Mining

Text mining, is the process of deriving knowledge / insights from textual data.
High-quality information is typically derived through the devising of patterns and
trends through pattern learning. Text mining involves the process of preprocessing
/ formatting the input text, finding patterns within the preprocessed data, and
finally evaluation of the results. The tasks of text mining are categorization of text,
clustering of text, concept / entity extraction, sentiment analysis, production of
granular taxonomies, entity relation modelling and document summarization.
This case study on text mining starts with using the twitter feeds from the dataset
“GameReview.csv” for further analysis. The extracted text is then transformed to
build a document-term matrix. Then, frequent words and associations are found
from the matrix. Important words in a document can be presented as a word cloud.
Packages used for text mining are “tm” and “wordcloud”.
> library(tm)
> dat<-read.csv(“GameReview.csv”,stringsAsFactors = FALSE)

> dat[1:5,4]
[1] “ SO ADDICTING DEFF DOWNLAOD ITS EPIC YOU CAT LOVERS WILL
FALL IN LOVE <3”
[2] “ Great game I love this game. Unlike other games they constantly give you money to
play. They are always given you a bone. Keep up the good work.”
[3] “ Sooo much FUN I would definitely recommend this game, it’s fun for dress up and
business. It’s extremely entertaining, I’m hooked already. :)”
[4] “ AWESOME Epic game so addictive 5stars <f0><U+009F><U+0098>
<U+0084>”
[5] “ Good game Great game”
The data is first converted to a corpus, which is a collection of text documents.

After that, the corpus can be processed with functions provided in package “tm”.
> corp <- Corpus(VectorSource(dat$text))
The corpus needs a couple of transformations, including changing letters to

lower case, and removing punctuations, numbers and stop words. The general
English stop-word list is tailored by adding and removing few words.
> corp <- tm_map(corp, content_transformer(tolower))
> corp <- tm_map(corp, removePunctuation)
> corp <- tm_map(corp, stripWhitespace)
> corp <- tm_map(corp, removeWords, stopwords(“english”))
In the above code, the function tm_map() is an interface to apply transformations

to the corpus. Word stemming can be done and after that, the stems can be
completed to their original forms, so that the words would look normal. This can be
achieved with function stemCompletion(). Before applying stemming a copy of the
corpus is taken to be used as dictionary for stem completion.
> corpcopy <- corp
> corp <- tm_map(corp, stemDocument)
> for (i in 1:5) {
235 Case Studies
> cat(paste(“[[“, i, “]] “, sep=””))

> writeLines(strwrap(corp[[i]], width=60))
>}
[[1]] addict deff downlaod epic cat lover will fall love 3
[[2]] great game love game unlik game constant give money play
alway given bone keep good work
[[3]] sooo much fun definit recommend game fun dress busi extrem
entertain im hook alreadi
[[4]] awesom epic game addict 5star f0u009fu0098u0084
[[5]] good game great game
After that, we use stemCompletion() to complete the stems with the unstemmed
corpus corpcopy as a dictionary. With the default setting, it takes the most frequent
match in dictionary as completion.
> stemCompletion2 <- function(x, dictionary) {
+ x <- unlist(strsplit(as.character(x), “ “))
+ x <- x[x != “”]
+ x <- stemCompletion(x, dictionary=dictionary)
+ x <- paste(x, sep=””, collapse=” “)
+ stripWhitespace(x)
+}
> corp <- lapply(corp, stemCompletion2, dictionary=corpcopy)
> corp <- Corpus(VectorSource(corp))
> for (i in 1:5) {
+ cat(paste(“[[“, i, “]] “, sep=””))
+ writeLines(strwrap(corp[[i]], width=60))
+ }
[[1]] addicted deff downlaod epic cat lovers will fall love 3
[[2]] great game love game unlike game constant give money play
always given bone keep good work

[[3]] sooo much fun definitely recommend game fun dress business
extremely entertained im hooked
[[4]] awesome epic game addicted 5stars f0u009fu0098u0084
[[5]] good game great game
From the listings from the corpus before stemming, after stemming and after
stemming completion we can see the words being changed. For example in line
4, the word was “addictive” before stemming, the word became “addict” after
stemming and then after stemming completion it became “addicted”.
The relationship between terms and documents are represented in a term-

document matrix, where each row stands for a term and each column stands for a
document. Each entry in this matrix is the number of occurrences of the term in
the document. Alternatively, one can also build a document-term matrix by swapping
row and column. We then build a term-document matrix from the above processed
corpus with function TermDocumentMatrix(). With its default setting, terms with
less than three characters are discarded.
> tdm <- TermDocumentMatrix(corp)
> tdm
<<TermDocumentMatrix (terms: 2117, documents: 1000)>>
Non-/sparse entries : 11144/2105856
Sparsity : 99%
Maximal term length : 391
Weighting : term frequency (tf)
As we can see from the above result, the term-document matrix is composed of
2117 terms and 1000 documents. It is very sparse, with 99% of the entries being zero.
We then have a look at the first six terms starting with “g” and tweets numbered
201 to 210.
> idx = grep(glob2rx(“g*”), dimnames(tdm)$Terms)
> inspect(tdm[idx,201:210])
<<TermDocumentMatrix (terms: 81, documents: 10)>>
237 Case Studies
Non-/sparse entries : 12/798

Sparsity : 99%
Maximal term length : 72
Weighting : term frequency (tf)
Sample :
Docs
Terms 201 202 203 204 205 206 207 208 209 210
game 1 2 1 1 0 0 0 1 2 0
gem 0 2 0 0 0 0 0 0 0 0
generally 0 0 0 0 0 0 0 0 0 0
get 0 1 0 0 0 0 0 2 0 0
give 0 0 0 0 0 0 0 0 0 0
given 0 0 0 0 0 0 0 0 0 0
glad 0 0 0 0 0 0 0 0 0 0
good 0 0 0 1 0 0 0 0 0 0
graphic 0 0 0 0 0 0 0 0 0 0
great 0 2 0 0 0 0 0 0 1 0
Many data mining tasks can be done, based on the above matrix. For example,
clustering, classification and association rule mining. When there are too many
terms, the size of a term-document matrix can be reduced by selecting terms that
appear in a minimum number of documents.
We will now have a look at the popular words and the association between
words from the 1000 tweets.
> findFreqTerms(tdm, lowfreq=80)
[1] “addicted” “love” “game” “good” “great” “play” “fun”
[8] “awesome” “get” “time” “like” “just” “update” “app”
[15] “can” “cant”
In the code above, findFreqTerms() finds frequent terms with frequency no less
than eighty. Note that they are ordered alphabetically, instead of by frequency or
popularity. To show the top frequent words visually, we next make a bar plot for them.
From the term document matrix, we can derive the frequency of terms with rowSums().
Then we select terms that appears in eighty or more documents and shown them with
a bar plot using the package ggplot2. In the code below, geom=”bar” specifies a bar
plot and coord_flip() swaps x-axis and y-axis. The bar plot below clearly shows that the
three most frequent words are “game”, “play” and “great”.
> termFrequency <- rowSums(as.matrix(tdm))
> termFrequency <- subset(termFrequency, termFrequency>=80)
> library(ggplot2)
> df <- data.frame(term=names(termFrequency), freq=termFrequency)
> ggplot(df, aes(x=term, y=freq)) + geom_bar(stat=”identity”) +
+ xlab(“Terms”) + ylab(“Count”) + coord_flip()
Figure 7.1 Ggplot of Terms and Frequencies
Alternatively, the above plot can also be drawn with barplot() as below, where
the argument las sets the direction of x-axis labels to be vertical.
> barplot(termFrequency, las=2)
239 Case Studies
Figure 7.2 Bar Plot of Terms and Frequencies
It is also possible to find the highly associated words with another word with
the function findAssocs(). Below is the code to find the terms associated with the
words “game” and “play” with correlation no less than 0.20 and 0.25 respectively.
The words are ordered by their correlation with the terms “game” (or “play”).
> findAssocs(tdm, “game”, 0.20)
$game
play player dont say new thatd ever
0.29 0.22 0.21 0.21 0.21 0.21 0.20
> findAssocs(tdm, “play”, 0.25)
$play
game potential course dont year will meter
0.29 0.27 0.27 0.26 0.26 0.25 0.25
After building a term-document matrix, the importance of words can be

shown with a word cloud, which can be easily produced with package wordcloud.
The code below, first converts the term-document matrix to a normal matrix,
and then calculates the word frequencies. After that, gray levels are set based on
word frequency and we use the function wordcloud() to make a plot for it. With
wordcloud(), the first two parameters give a list of words and their frequencies.
Words with frequency below twenty are not plotted, as specified by min.freq=20. By
setting random.order=F, frequent words are plotted first, which makes them appear
in the centre of cloud. We also set the colours to gray levels based on frequency.
> library(wordcloud)
> m <- as.matrix(tdm)
> wordFreq <- sort(rowSums(m), decreasing=TRUE)
> pal <- brewer.pal(9, “BuGn”)
> pal <- pal[-(1:4)]
> set.seed(375)
> grayLevels <- gray( (wordFreq+10) / (max(wordFreq)+10) )
> wordcloud(words=names(wordFreq), freq=wordFreq, min.freq=20,
random.order=F, colors=pal)
Figure 7.3 Word Cloud based on Frequencies
The above word cloud clearly shows again that “game”, “play” and “great” are
the top three words, which validates that the Game review tweets. Some other
important words are “love”, “good” and “fun”, which shows that the review on the
game is very good.
With hierarchical clustering, we find clusters of words. The plot of clustering

should not be crowded with words and hence we remove the sparse terms. Then
241 Case Studies
the distances between terms are calculated with dist() after scaling. After that, the
terms are clustered with hclust() and the dendrogram is cut into 4 clusters. The
agglomeration method is set to ward, which denotes the increase in variance when
two clusters are merged.
> tdm2 <- removeSparseTerms(tdm, sparse=0.90)
> m2 <- as.matrix(tdm2)
> distMatrix <- dist(scale(m2))
> fit <- hclust(distMatrix, method=”ward.D”)
> plot(fit)
> rect.hclust(fit, k=4)
> (groups <- cutree(fit, k=4))
addicted love game good great play fun awesome get
1 1 2 3 4 1 4 1 1
time
1
Figure 7.4 Cluster of Words with High Frequencies Using Hierarchical Clustering
In the above dendrogram, we can see the topics in the tweets grouped into 4
different clusters. The most frequent word “game” is in the second cluster, the next
frequent word “good” is in the third cluster, the words “great” and “fun” falls under the
fourth cluster and the remaining words with low frequency falls under the first cluster.
Next we cluster the tweets using k-means clustering algorithm. The k-means
clustering takes the values in the matrix as numeric. We transpose the term-document
matrix to a document-term one. The tweets are then clustered with kmeans() with
the number of clusters set to eight. After that, we check the popular words in every
cluster and also the cluster centres. A fixed random seed is set with set.seed() before
running kmeans().
> m3 <- t(m2)
> set.seed(123)
> k <- 8
> kmeansResult <- kmeans(m3, k)
> round(kmeansResult$centers, digits=3)
addicted love game good great play fun awesome get time
1 0.168 0.107 1.267 0.137 1.313 0.069 0.176 0.130 0.145 0.061
2 0.227 0.394 3.652 0.409 0.348 0.803 0.273 0.258 0.409 0.258
3 0.038 0.017 1.174 0.547 0.000 0.127 0.068 0.127 0.081 0.106
4 0.212 0.061 1.091 0.333 0.030 0.606 0.212 0.030 2.242 0.788
5 0.183 0.113 1.507 0.113 0.183 0.887 1.085 0.197 0.197 0.254
6 1.205 0.154 0.397 0.077 0.090 0.154 0.423 0.179 0.026 0.064
7 0.000 0.118 0.000 0.125 0.104 0.146 0.212 0.090 0.066 0.097
8 0.237 1.381 1.216 0.031 0.031 0.216 0.062 0.175 0.134 0.093
To make it easy to find what the clusters are about, we then check the top three
words in every cluster.
> for (i in 1:k) {
+ cat(paste(“cluster “, i, “: “, sep=””))
+ s <- sort(kmeansResult$centers[i,], decreasing=T)
+ cat(names(s)[1:3], “\n”)
+}
243 Case Studies
cluster 1: great game fun

cluster 2: game play good
cluster 3: game good play
cluster 4: get game time
cluster 5: game fun play
cluster 6: addicted fun game
cluster 7: fun play good
cluster 8: love game addicted
From the above top words and centres of clusters, we can see that the clusters are
of different topics. We can see in every cluster except the cluster 7, the word “game” is
a part of it and each of these clusters talk about the game in different angle.
7.2. Credit Risk Analysis and Prediction

Credit risk evaluation has become more important nowadays for Banks to issue
loans for their customers based on their credibility. For this the internal rating based
approach is the most sought by the banks that need approval by the bank manager.
The most accurate and highly used credit scoring measure is the Probability of
Default called the PD. Defaulter is the one who is unlikely to repay the loan amount
or will have overdue of loan payment by more than 90 days. Hence determining the
PD is the crucial step for credit scoring of the customers seeking bank loan.
Hence a data mining framework is developed for PD estimation from a given

set of data using the data mining techniques available in R Package. The data used
to implement and test this model is taken from the UCI Repository. The German
credit scoring dataset with 1000 records and 21 attributes is used for this purpose.
The numeric format of the data is loaded into the R Software and a set of data
preparation steps are executed before the same is used to build the classification
model. The dataset that we have selected does not have any missing data. But, in
real time there is possibility that the dataset has many missing or imputed data
which needs to be replaced with valid data generated by making use of the available
complete data. The k nearest neighbours algorithm is used for this purpose to
perform multiple imputation. This is implemented using the knnImputation()
function of package DMwR. The numeric features are normalized before this step.
The dataset has many attributes that define the credibility of the customers
seeking for several types of loan. The values for these attributes can have outliers
that do not fit into the regular range of data. Hence, it is required to remove the
outliers before the dataset is used for further modelling. The outlier detection for
quantitative features is done using the function levels(). For numeric features the
box plot technique is used for outlier detection and this is implemented using the
daisy() function of the cluster package. But, before this the numeric data has to
be normalized into a domain of [0, 1]. The agglomerative hierarchical clustering
algorithm is used for outlier ranking. This is done using the outliers.ranking()
function of the DMwR package. After ranking the outlier data, the ones that is out
of range is disregarded and the remaining outliers are filled with null values.
The inconsistencies in the data like unbalanced dataset have to be balanced
before building the classification model. Many real time datasets have this problem
and hence need to be rectified for better results. But, before this step, it is required
to split the sample dataset into training and test datasets which will be in the ratio
4:1 (i.e. Training dataset 80% of data and 20% of data will be test dataset). Now
the balancing step will be executed on the training dataset using the SMOTE()
function of the DMwR package.
Next using the training dataset the correlation between the various attributes
need to be checked to see if there are any redundant information represented using
two attributes. This is implemented using the plotcorr() function in the ellipse
package. The unique features will then be ranked and based on the threshold limit
the number of highly ranked features will be chosen for model building. For ranking
the features the randomForest() function of the randomForest package is used. The
threshold for selecting the number of important features is chosen by using the
rfcv() function of the randomForest package.
Now the resultant dataset with the reduced number of features is ready for use
by the classification algorithms. Classification is one of the data analysis methods
that predict the class labels. Classification can be done in several ways and one of
the most appropriate for the chosen problem is using decision trees. Classification
is done in two steps – (i) the class labels of the training dataset is used to build the
decision tree model and (ii) This model will be applied on the test dataset to predict
the class labels of the test dataset. For the first step the function rpart() of the rpart
package will be used. The predict() function is used to execute the second step. The
resultant prediction is then evaluated against the original class labels of the test
dataset to find the accuracy of the model.
Dataset Attribute Types
A1 A2 A3 A4 A5 A6 A7 A8 A9 A10 A11 A12 A13 A14 A15 A16 A17 A18 A19 A20 Def
Q N Q Q N Q Q N Q Q N Q N Q Q N Q N B B B
Q: Quantitative N: Numeric B: Binary
A1: Status of Existing Account

(1: < 0 DM, 2: < 200 DM, 3: >= 200 DM, 4: No existing Account)
A2: Loan Duration in Month
A3: Credit History
(0: No credits taken so far, 1: All credit in this Bank paid back duly, 2: Existing credits paid back dully till
now,
245
3: Delay in paying off in the past, 4: Credits existing in other banks)

A4: Loan Purpose
(0: new car purchase, 1: used car purchase, 2: furniture or equipment purchase, 3: radio or television
purchase, 4: domestic appliances purchase, 5: repairs, 6: education, 7: vacation, 8: retraining, 9: Business,
10: others)
Case Studies
A5: Credit Amount (in DM)

A6: Bonds / Savings
(1: < 100 DM, 2: >= 100 and < 500 DM, 3: >= 500 DM and 1000 DM, 4: >= 1000 DM, 5: no savings /
bonds)
A7: Present Employment Since
(1: unemployed, 2: < 1 year, 3: >= 1 and < 4 years, 4: >= 4 and < 7 years, 5: >= 7 years)
A8: Instalment rate in percentage of disposable income
A9: Personal Status and Sex
(1: Divorced Male, 2: Divorced/Married Female, 3: Male Single, 4: Married Male, 5: Female Single)
A10: Other Debtors / Guarantors
(1: None, 2: Co-applicant, 3; Guarantor)
A11: Present Residence Since (in Years)
A12: Property
(1: Real Estate, 2: Life Insurance, 3: Car or others, 4: No property)
A13: Age in years
A14: Other instalment plans
(1: Bank, 2: Stores, 3: None)
A15: Housing
(1: Rented, 2: Owned, 3: For Free)
A16: Number of existing credits at this bank
A17: Job Status
(1: Unemployed non-resident, 2: Unemployed resident, 3: Skilled Employee, 4: Self-Employed )
A18: Number of People being liable to provide maintenance for
A19: Telephone
(0: Not Available, 1: Available)
A20: Foreign Worker
(0: No, 1: Yes)
Def: Class Label
(0: Non Default, 1: Default)
247 Case Studies
Dataset Selection
The German Credit Scoring dataset in the numeric format which is used for the
implementation of this model has the below attributes and the descriptions of the
same are given in the below table.
After selecting and understanding the dataset it is loaded into the R software
using the below code. The dataset is loaded into R with the name creditdata.
> creditdata <- read.csv(“UCI German Credit Data Numeric.csv”,
header = TRUE, sep = “,”)
> nrow(creditdata)
[1] 1000
Data Pre-Processing
1) Outlier Detection: To identify the outliers of the numeric attributes, the
values of the numeric attributes are normalized into the domain range of [0,
1] and they are plotted as box plot to view the outlier values as in Fig. 7.5. The
code and the result for this step are given as below.
> normalization <- function(data,x)
> {for(j in x)
> {data[!(is.na(data[,j])),j]=
> (data[!(is.na(data[,j])),j]-min(data[!(is.na(data[,j])),j]))/
> (max(data[!(is.na(data[,j])),j])-min(data[!(is.na(data[,j])),j]))}
> return(data)}
> c <- c(2,5,8,11,13,16,18)
> normdata <- normalization(creditdata,c)
> boxplot(normdata[,c])
Figure 7.5 Box Plot of Outliers in Numeric Attributes
To identify the outliers of the quantitative attributes, the below commands

are used. From the results of the same, one can identify the values that do not fall
under the allowed values. For example the attribute “A1” can only take the values
“1”, “2”, “3”, “4”. If there is any observation that has data other than these allowed
values, it is removed. Similarly, the allowed values for each quantitative attribute
can be checked and outliers removed.
> levels(as.factor(creditdata[,”A1”]))
[1] “1” “2” “3” “4”
2) Outliers Ranking: The agglomerative hierarchical clustering algorithm
chosen for ranking the outliers is less complex and easy to understand. Each
observation is assumed to be a cluster and in each step, the observations
are grouped based on the distance between them. Each observation that is
observed later has lower rank. It is seen that the observations with lower rank
are outliers because there are dissimilarities between them and the other
observations. For outlier ranking the following code is used.
> require(cluster)
> distance=daisy(creditdata[,-19],stand=TRUE,metric=c(“gower”),
type = list(interval=c(2,5,8,11,13,16,18),
nominal=c(1,3,4,6,7,9,10,12,14,15,17),binary=c(19,20)))
> require(DMwR)
249 Case Studies
> outlierdata=outliers.ranking(distance,test.data=NULL,method=”sizeDiff ”,
clus = list(dist=”euclidean”, alg = “hclust”, meth=”average”),
power = 1, verb = F)
3) Outliers Removal: The observations which are out of range (based on the
rankings) are removed using the below code. After outlier removal the dataset
creditdata is renamed as creditdata_noout.
> boxplot(outlierdata$prob.outliers[outlierdata$rank.outliers])
> n=quantile(outlierdata$rank.outliers)
> n1=n[1]
> n4=n[4]
> filler=(outlierdata$rank.outlier > n4*1.3)
> creditdata_noout=creditdata[!filler,]
> nrow(creditdata_noout)
[1] 975
4) Imputations Removal: The method used for null values removal is multiple
imputation method in which the k nearest neighbours’ algorithm is used for
both numeric and quantitative attributes. The numeric features are normalized
before calculating the distance between objects. The following code is used for
imputations removal. After imputations removal the dataset creditdata_noout is
renamed as creditdata_noout_noimp.
> require(DMwR)
> creditdata_noout_noimp=knnImputation(creditdata_noout, k = 5, scale = T,
meth = “weighAvg”, distData = NULL)
> nrow(creditdata_noout_noimp)
[1] 975
There were no null values for the attributes in the dataset we have chosen and
hence the number of records remains unchanged after the above step.
5) Splitting Training and Test Datasets: Before proceeding to the further steps,
the dataset has to be split into training and test datasets so that the model can
be built using the training dataset. The code for splitting the database is listed
below.
> library(DMwR)
> split<-sample(nrow(creditdata_noout_noimp),
round(nrow(creditdata_noout_noimp)*0.8))
> trainingdata=creditdata_noout_noimp[split,]
> testdata=creditdata_noout_noimp[-split,]
6) Balancing Training Dataset: The SMOTE() function handles unbalanced
classification problems and it generates the new smoted dataset that addresses
the unbalanced class problem. It artificially generates observations of minority
classes using the nearest neighbours of this class of elements to balance the
training dataset. The following code is used for balancing the training dataset.
> creditdata_noout_noimp_train=trainingdata
> creditdata_noout_noimp_train$default <-
factor(ifelse(creditdata_noout_noimp_train$Def == 1, “def ”, “nondef ”))
> creditdata_noout_noimp_train_smot <-
SMOTE(default ~ ., creditdata_noout_noimp_train, k=5,perc.over = 500)
Figure 7.6 Data Distribution before Balancing

251 Case Studies
Figure 7.7 Data Distribution after Balancing
The data distribution before and after balancing the data are shown in the
figures Fig. 7.6 and Fig. 7.7. This method is based on proximities between objects
and produces a spatial representation of these objects. Proximities represent the
similarity or dissimilarity between data objects. The code used to plot these objects
is shown below.
> library(cluster)
> dist1=daisy(creditdata_noout_noimp_train[,-21],stand=TRUE,metric=c(“gower”),
> dist2=daisy(creditdata_noout_noimp_train_smot[,-21],
stand=TRUE,metric=c(“gower”),
> loc1=cmdscale(dist1,k=2)
> loc2=cmdscale(dist2,k=2)
> x1=loc1[,1]
> y1=loc1[,2]
> x2=loc2[,1]
> y2=loc2[,2]
> plot(x1,y1,type=”n”)
> text(x1,y1,labels=creditdata_noout_noimp_train[,22],
col=as.numeric(creditdata_noout_noimp_train[,22])+4)
> plot(x2,y2,type=”n”)
> text(x2,y2,labels=creditdata_noout_noimp_train_smot[,22],
col=as.numeric(creditdata_noout_noimp_train_smot[,22])+4)
Features Selection
1) Correlation Analysis: Datasets may contain irrelevant or redundant features
which might make the model more complicated. Hence removing such redundant
features will speed up the model. The function plotcorr() plots a correlation matrix
using ellipse shaped glyphs for each entry. It shows the correlation between the
features in an easy way. The plot is coloured for more clarity. The following code
displays the correlation. Correlation is checked independently for each data type:
numeric and nominal. From the results in the figures, Fig. 7.8 and Fig. 7.9, it is
observed that there is no positive correlation between any of the features, both
numeric and quantitative. Hence, in this step none of the features are removed.
> library(ellipse)
> c= c(2,5,8,11,13,16,18)
> plotcorr(cor(creditdata_noout_noimp_train[,c]),col=cl<-c(7,6,3))
> c= c(1,3,4,6,7,9,10,12,14,15,17)
> plotcorr(cor(creditdata_noout_noimp_train [,c]),col=cl<-c(“green”,”red”,”blue”))
Figure 7.8 Correlation of Numeric Features

253 Case Studies
Figure 7.9 Correlation of Quantitative Features

2) Ranking Features: The aim of this step is to find the subset of features that
will be really relevant for the analysis as irrelevant features causes drawbacks like
increased runtime, complex patterns etc. This resultant subset of features should
give the same results as that of the original dataset. The proposed method picks a
random object from the observations and generates several trees and on the basis
of the accuracy of classifier or error ratio, features are weighted. To make the
table of important features the following code is used.
> set.seed(454)
> data.frame(creditdata_noout_noimp_train)
> randf<-randomForest(Def~ ., data=creditdata_noout_noimp_train, ntree=700,
importance=TRUE, proximity=TRUE)
> importance(randf, type=1, scale=TRUE)
The above function importance() displays the features importance using the
“mean decrease accuracy” measure in the below table. The measures can be plotted
using the function varImpPlot() as shown in the figure. Fig. 7.10.
Features Mean Decrease Accuracy

A1 8.085083
A2 7.070556
A3 4.691744
A4 -0.10716
Features Mean Decrease Accuracy
A5 6.238347
A6 4.554283
A7 3.316346
A8 0.59622
A9 1.634721
A10 1.383725
A11 0.541585
A12 2.344433
A13 2.621854
A14 4.629331
A15 0.825801
A16 1.225997
A17 0.635881
A18 0.037408
A19 1.117891
A20 1.388876
> varImpPlot(randf)
Figure 7.10 Plot Showing Importance of Features

255 Case Studies
3) Features Selection: To fix the number of features to be selected based on the

ranking, a threshold is required. This is accomplished using the below code.
> findopt=rfcv(creditdata_noout_noimp_train[,-21],
> creditdata_noout_noimp_train[,21], cv.fold=10, scale=”log”, step=0.9)
> opt <- which.max(findopt$error.cv)
> plot( findopt$n.var, findopt$error.cv, type= “h”, main = “Importance”,
xlab=”Number of Features”, ylab = “Classifier Error Rate”)
> axis(1, opt, paste(“Threshold”, opt, sep=”\n”), col = “red”, col.axis = “red”)
The result of this code is shown in the figure Fig. 7.11 and it shows the best
number of features is 15. Hence we select the features A1, A2, A3, A5, A6, A7, A9,
A10, A12, A13, A14, A16, A19, A20, Def to build the model.
Building Model
Classification is one of the data analysis forms that predicts categorical labels. We
used the decision tree model to predict the probability of default. The following
code uses the function rpart() and finds a model from the training dataset.
> library(rpart)
> c = c(4, 8, 11, 15, 17, 18, 22)
> trdata=data.frame(creditdata_noout_noimp_train[,-c])
> tree=rpart(trdata$Def~.,data=trdata,method=”class”)
> printcp(tree)
Figure 7.11 Threshold for Choosing the Number of Features

The result of this code is displayed below and in the table below.
Classification tree:
rpart(formula = trdata$Def ~ ., data = trdata, method = “class”)
Variables actually used in tree construction:
[1] A1 A12 A13 A2 A3 A5 A6 A9
Root node error: 232/780 = 0.29744
n= 780
CP nsplit rel error Xerror xstd

1 0.049569 0 1 1 0.05503
2 0.012931 4 0.78448 0.84483 0.052215
3 0.011494 5 0.77155 0.88793 0.053071
4 0.010057 9 0.72414 0.89655 0.053235
5 0.01 18 0.61207 0.89655 0.053235
The command to plot the classification tree is shown in the figure Fig. 7.12.
> plot(tree, uniform=TRUE,main=”Classification Tree”)
> text(tree, use.n=TRUE, all=TRUE, cex=0.7)
Figure 7.12 Classification Tree Model

257 Case Studies
Prediction
The model is tested using the test dataset by using the predict() function. The code
for the same and the results of the prediction are displayed below and in the table
below.
> predicttest=data.frame(testdata)
> pred=predict(tree,predicttest)
> c=c(21)
> table(predict(tree, testdata, type=”class”,na.action=na.pass), testdata[, c])
def Nondef
def 30 5
nondef 6 154
Evaluation
Common metrics calculated from the confusion matrix are Precision, Accuracy,
True Positive Rate (TP Rate) and False Positive Rate (FP Rate). The calculations
for the same are listed below.
True Defaults
Precision =
True Defaults + False Defaults
True Defaults + True Non defaults

Accuracy =
Total Test set
True Defaults
TP Rate =
Total Defaults
False Defaults
FP Rate =
Total Non defaults
From our resultant data we get the values of the above metrics by applying the
values as derived below.
True Defaults = 30
False Default = 6
Total Default = 35
True Non default = 154
False Non default = 5
Total Non default = 160
Total Test set = 195
Precision = 30 / (30 + 6) = 0.833
Accuracy = (30 + 154) / 195 = 0.943
TP Rate = 30 / 35 = 0.857
FP Rate = 6 / 160 = 0.037
Precision Accuracy TP Rate FP Rate

0.833 0.943 0.857 0.037
These results show that the proposed model is performing with high accuracy
and precision and hence can be applied for credit scoring.
7.3. Exploratory Data Analysis on Crime Data

This case study is to do exploratory data analysis by inspecting, cleansing, transforming
and find meaningful insights of the publicly available crime data of US. The goal
is to discover useful information, suggest conclusion and support decision making.
The dataset is loaded into the R tool and understood completely. The unwanted
data are removed, that is the data is being pre-processed. As outcome the crime
which leads to arrest, frequency of different crime types, when crime takes place
are being found out. The result that is being found out is being visualized using bar
plots and heat maps. As the results are visualized it is easy to draw conclusions. The
objective is to identify criminal hot spots dynamically, to facilitate precautionary
259 Case Studies
deployment of police resources, to perform temporal data analysis of crime data and
to capture the trend of crimes happening.
This case study will have the below functions:

1. Gather and load the US Crime data into R Tool
a. Understand the data field organization and its important dimensions
b. Each crime incident indicates one record
c. Find the different crime types
2. Pre-process the data
a. Removing duplicate records
b. Removing records with missing values
c. Removing records with incorrect values
d. Formatting the Timestamp field (splitting the date and time parts of
the data)
e. Binning the time intervals (4 intervals – each 6 hours)
f. Grouping of similar crimes into one crime type
3. Exploratory analysis of the real time crime data
a. Find and visualize which crime types lead to arrests
b. Find and visualize frequency of different crime types
c. Find and visualize the hours of the day in which more crimes happen
d. Find and visualize the days of a week in which more crimes happen
e. Find and visualize the months of a year in which more crimes occur
f. Visualize which crime type happens more during which hours of the day
g. Visualize which crime type happens more during which day of the
week
h. Visualize which crime type happens more during which month of the
year
The packages used in this case study are:
chron - Creates chronological objects which represent dates and time of a day.
ggplot2 - A system for creating graphs, based on the data you provide.
The function used are:
read.csv() - Reads a file in table format and creates a data frame from it.
subset() - Return subsets of vectors, matrices or data frames which meet

conditions.
is.na() - NA is a logical constant of length 1 which contains a missing value

indicator.
which() - Gives the TRUE indices of a logical object, allowing for array indices
as.POSIXlt() - Function to manipulate objects of class “POSIXlt” representing

calendar dates and times.
times() - Creates objects which represent date or time.
format() - Format an R object for pretty printing.
cut() - Cut divides the range of x into intervals and codes the values in x according
to which interval they fall. The leftmost interval corresponds to level one, the next
leftmost to level two and so on.
labels() - Finds a suitable set of labels from an object for use in printing or
plotting
table() - Table uses the cross-classifying factors to build a contingency table of

the counts at each combination of factor levels.
strptime() - Function to convert between character representations and objects

of class “POSIXlt” representing calendar dates and times.
weekdays() - Extracts the day of the week
months() - Extracts the month
length() - Get or set the length of vectors (including lists) and factors, and of
any other R object for which a method has been defined.
261 Case Studies
unique() - Removes duplicated elements/rows from a vector, data frame or array
as.character() - Creates or tests for objects of type “character”
ifelse() - Ifelse returns a value with the same shape as test which is filled
with elements selected from either yes or no depending on whether the element
of test is TRUE or FALSE.
factor() - The function factor is used to encode a vector as a factor

(the terms ‘category’ and ‘enumerated type’ are also used for factors). If
argument ordered is TRUE, the factor levels are assumed to be ordered.
aggregate() - Splits the data into subsets, computes summary statistics for each,
and returns the result in a convenient form.
qplot() - It’s is the basic plotting function. It is used to create a different number
of plots.
ggplot() - It initializes a ggplot object. It can be used to declare input data

frame for a graph and to specify the set of plot aesthetics.
The dataset was analyzed to get details such as the file size, number of records,
the fields specified and their meaning.
Dataset Size No. of No. of Attribute Names

Name Attributes Records
crime.csv 44.9 17 262601 Case, Date of occurrence, Block,
MB IUCR,
Primary description, Secondary
description,
Location description, Arrest,
Domestic,
Beat, Ward, FBI.CD,
X coordinate, Y coordinate,
Latitude, Longitude, Location
U.S crime dataset is loaded into the R tool and the data field organization
and its important dimensions are understood. It is noted that each crime incident
indicates one record and the various crime types are manually analyzed. The dataset
is loaded using the read.csv() function. After this the required packages for this
project are installed and loaded using the below commands.
> install.packages(“chron”)
> library(chron)
> install.packages(“Rcpp”, dependencies = TRUE)
> library(Rcpp)
> install.packages(“ggplot2”)
> library(ggplot2)
The pre-processing deals with removing duplicate records, removing records with
missing values, removing records with incorrect values, formatting the Timestamp
field (splitting the date and time parts of the data), binning the time intervals (4
intervals – each 6 hours) and grouping of similar crimes into one crime type. The
functions used for these preprocessing steps are subset(), as.POSIXlt(), weekdays(),
months(), chron(), cut(), table(), length(), as.character(), ifelse().
Then finally we go for finding and visualizing which crime types lead to arrests,
finding and visualizing frequency of different crime types, finding and visualizing
the hours of the day in which more crimes happen, finding and visualizing the days
of a week in which more crimes happen, finding and visualizing the months of a
year in which more crimes occurs, visualizing the occurrence of various crime types
during various hours of the day, visualizing the occurrence of various crime types
during various days of the week, visualizing the occurrence of various crime types
during various months of the year. All these exploration and visualization are done
using the functions qplot(), factor(), aggregate() and ggplot() that belong to the
package ggplot2.
The code for the entire case study is as follows.

> setwd(“F:/”)
#Loading data
> crime.data <- read.csv(“crime.data.csv”)
263 Case Studies
#Removing duplicate records

> crime.data <- subset(crime.data, !duplicated(crime.data$CASE.))
> crime.data <- subset(crime.data, !is.na(crime.data$LATITUDE))
> crime.data <- subset(crime.data, !is.na(crime.data$WARD))
#Removing incorrect records

> crime.data <- crime.data[crime.data$CASE.!= “CASE#”,]
#Date Conversion
> crime.data$date <- as.POSIXlt(crime.data$DATE..OF.OCCURRENCE,
format= “%m/%d/%Y %I:%M:%S %p”)
> crime.data <- subset(crime.data, !is.na(crime.data$date))
> crime.data$time <- times(format(crime.data$date, “%H:%M:%S”))
#Binning Time Intervals

> time.tag <- chron(times = c(“00:00:00”, “06:00:00”, “12:00:00”, “18:00:00”,
“23:59:00”))
> crime.data$time.tag <- cut(crime.data$time, breaks = time.tag,
labels = c(“00-06”,”06-12”, “12-18”, “18-00”), include.lowest = TRUE)
> table(crime.data$time.tag)
#Formating Date
> crime.data$date <- as.POSIXlt(strptime(crime.data$date, format = “%Y-%m-%d”))
> crime.data$day <- weekdays(crime.data$date, abbreviate = TRUE)
> crime.data$month <- months(crime.data$date, abbreviate = TRUE)
#Finding unique crime types

> table(crime.data$PRIMARY.DESCRIPTION)
> length(unique(crime.data$PRIMARY.DESCRIPTION))
> crime.data$crime <- as.character(crime.data$PRIMARY.DESCRIPTION)
#Grouping crime types

> crime.data$crime <- ifelse(crime.data$crime %in% c(“CRIM SEXUAL ASSAULT”,
“PROSTITUTION”, “SEX OFFENSE”), “SEX”, crime.data$crime)
> crime.data$crime <- ifelse(crime.data$crime %in% c(“MOTOR VEHICLE
THEFT”), “MVT”, crime.data$crime)
> crime.data$crime <- ifelse(crime.data$crime %in%
c(“GAMBLING”,INTERFERE WITH PUBLIC OFFICER”, “INTERFERENCE
WITH PUBLIC OFFICER”, “INTIMIDATION”, “LIQUOR LAW
VIOLATION”, “OBSCENITY”, “NON CRIMINAL”, “PUBLIC PEACE VIOLATION”,
“PUBLIC INDECENCY”, “STALKING”, “NON-CRIMINAL (SUBJECT
SPECIFIED)”), “NONVIO”, crime.data$crime)
> crime.data$crime <- ifelse(crime.data$crime == “CRIMINAL DAMAGE”,
“DAMAGE”, crime.data$crime)
> crime.data$crime <- ifelse(crime.data$crime == “CRIMINAL
TRESPASS”,”TRESPASS”, crime.data$crime)
> crime.data$crime <- ifelse(crime.data$crime %in% c(“NARCOTICS”,
“OTHERNARCOTIC VIOLATION”, “OTHER NARCOTIC VIOLATION”),
“DRUG”, crime.data$crime)
> crime.data$crime <- ifelse(crime.data$crime == “DECEPTIVE
PRACTICE”,”FRAUD”, crime.data$crime)
> crime.data$crime <- ifelse(crime.data$crime %in% c(“OTHER OFFENSE”,
“OTHEROFFENSE”), “OTHER”, crime.data$crime)
> crime.data$crime <- ifelse(crime.data$crime %in% c(“KIDNAPPING”,
“WEAPONSVIOLATION”, “OFFENSE INVOLVING CHILDREN”), “VIO”,
crime.data$crime)
> table(crime.data$crime)
> length(unique(crime.data$crime))
#Finding crimes that leads to arrest

> crime.data$ARREST <- ifelse(as.character(crime.data$ARREST) == “Y”, 1, 0)
> crime.data.arrest <- crime.data[crime.data$ARREST != 0,]
265 Case Studies
> table(crime.data.arrest$crime)
#Visualizing Crimes that leads to arrest

> qplot(crime.data.arrest$crime, xlab = “Crime”, main = “Crimes in Chicago that
leads to
arrest”) + scale_y_continuous(“Number of crimes”) + theme(text = element_
text(size=10),
axis.text.x = element_text(angle=90, hjust=1))
#Visualizing Various Crime Types

> qplot(crime.data$crime, xlab = “Crime”, main = “Crimes in Chicago”) +
scale_y_continuous(“Number of crimes”) + theme(text = element_text(size=10),
axis.text.x = element_text(angle=90, hjust=1))
#Visualizing Various Crimes by time of day

> qplot(crime.data$time.tag, xlab=”Time of day”, main = “Crimes by time of day”) +
scale_y_continuous(“Number of crimes”
#Visualizing Various Crimes by day of week
> crime.data$day <- factor(crime.data$day,
levels = c(“Mon”, “Tue”, “Wed”, “Thu”, “Fri”, “Sat”, “Sun”))
> qplot(crime.data$day, xlab = “Day of week”, main= “Crimes by day of week”) +
scale_y_continuous(“Number of crimes”)
#Visualizing Various Crimes by month

> crime.data$month <- factor(crime.data$month, levels = c(“Jan”, “Feb”, “Mar”,
“Apr”, “May”, “Jun”, “Jul”, “Aug”, “Sep”, “Oct”, “Nov”, “Dec”))
> qplot(crime.data$month, xlab = “Month”, main= “Crimes by month”) +
scale_y_continuous(“Number of crimes”)
#Visualizing Crimes by Time of Day using Heatmap

> temp <- aggregate(crime.data$crime, by= list(crime.data$crime,crime.data$time.
tag), FUN= length)
> names(temp) <- c(“crime”, “time.tag”, “count”)

> ggplot(temp, aes(x = crime, y = factor(time.tag))) + geom_tile(aes(fill = count)) +
scale_x_discrete(“Crime”, expand = c(0,0)) +
scale_y_discrete(“Time of day”, expand = c(0,-2)) +
scale_fill_gradient(“Number of crimes”, low = “white”, high = “steelblue”) +
theme_bw() + ggtitle(“Crimes by time of day”) +
theme(panel.grid.major = element_line(colour = NA),
panel.grid.minor = element_line (colour = NA)) +
theme(text = element_text(size=10), axis.text.x = element_text(angle=90, hjust=1))
#Visualizing Crimes by Day of Week using Heatmap

> temp <- aggregate(CASE. ~ crime + day, data = crime.data, FUN = length)
> names(temp)[3] <- “count”
> ggplot(temp, aes(x = crime, y = day, fill = count)) + geom_tile(aes(fill = count)) +
scale_y_discrete(“Day of week”, expand = c(0,-2)) +
theme_bw() + ggtitle(“Crimes by day of week”) +
theme(panel.grid.major = element_line(colour = NA),
panel.grid.minor = element_line(colour = NA)) +
#Visualizing Crimes by Month of Year using Heatmap

> temp <- aggregate(CASE. ~ crime + month, data = crime.data, FUN = length)
> names(temp)[3] <- “count”
> ggplot(temp, aes(x = crime, y = month, fill = count)) + geom_tile(aes(fill = count)) +
scale_y_discrete(“Month”, expand = c(0,-2)) +
267 Case Studies

theme_bw() + ggtitle(“Crimes by Month”) +
theme(panel.grid.major = element_line (colour = NA),
panel.grid.minor = element_line(colour = NA)) +
theme(text = element_text(size=10), axis.text.x = element_text(angle=90, hjust=1)) +
Figure 7.13 Ggplot of Crimes Leading to Arrest
Figure 7.14 Frequencies of Crimes

Figure 7.15 Ggplot of Crimes by Time of the Day
Figure 7.16 Ggplot of Crimes by Day of the Week

269 Case Studies
Figure 7.17 Ggplot of Crimes by Month
Figure 7.18 Heatmap of Crimes by Time of the Day

Figure 7.19 Heatmap of Crimes by Day of the Week
Figure 7.20 Heatmap of Crimes by Month of the Year
7.4. Exploratory Data Analysis on Baseball Data

This case study is to find meaningful information by visualization about the Base
Ball Teams by using the packages and functions in R. Data is spread across three
files, namely, salaries.csv, master.csv and batting.csv. These datasets are loaded and
sequences of pre-processing are done. The pre-processed data is then analyzed
271 Case Studies
on several aspects such as variation of salary over a period of time with respect to
League, Team etc. The player’s career records are analyzed based on their hits and
runs and their batting averages calculated. These analysis results are then presented
in the form of graphs and histograms using the functions available in R. The
objective is to identify the trend of base ball players salary over the years, facilitates
to understand the correlation between players salary and their performances, analyze
if age, country, height and weight of the players have impact on their performance
and it captures the details of top performing baseball players.
This case study will have the below functions:

1. Gather and load the US Baseball data into R Tool
a. Understand the data field organization and its important dimensions
b. The data is split among three files namely Master.csv, Batting.csv and
Salaries.csv
2. Pre-process the data
a. Filtering records specific to a particular periods and specific leagues
b. Sorting the filtered data based on the chronological order of records
c. Summarising data to get min, max and average salary year wise
d. Summarising data to get min, max and average salary league wise
e. Summarising data to get min, max and average salary team wise
f. Summarising to get min, max and average salary year wise and league
wise
g. Summarising to get min, max and average salary year wise and team
wise
h. Merging the salary data of the base ball players with their biographical
(personal) data – (To combine player’s last and first name into single
name)
i. Merging the salary data with batting statistics (history data) data
based on the year, player, Team and League fields (To study salary
given based on performance)
j. Merging the salary, master and batting datasets (To study player wise
batting performance)
k. Filter the records of the players in the years in which they have not
had a chance to bat (AB > 0)
c. Exploratory analysis of the baseball team data
a. Visualize the trend of how salaries change over time
b. Find one players salary, team and other details
c. Find the relation of the player’s salary with his height, weight, birth
country
d. Find how each player was batting year wise
e. Visualize correlation of salary with the players performance
f. Visualize each player’s career record (Eg. Total Hit and Runs) based
on their highest rank
g. Visualize the correlation between the players hits and runs
h. Visualize the batting average of the players in a histogram
The packages used in this case study are:
data.table - Extension of package “data.frame”. Does fast aggregation of large

data, ordered joins, add/modify/delete of columns by group, a fast friendly file reader
and parallel file writer.
ggplot2 - A system for creating graphs, based on the data you provide.
The function used are:
read.csv() - Reads a file in table format and creates a data frame from it.
as.data.table() - It is a function to check if an object is of type “data.table”.
merge() - It is used to merge two data frames by common column or row names,
or do other versions of database join operations.
order() - Order returns a permutation which rearranges its first argument into
ascending or descending order, breaking ties by further arguments.
max() - Returns the (parallel) maxima and minima of the input values.
min() - Returns the (parallel) maxima and minima of the input values.
273 Case Studies
mean() - Generic function for the (trimmed) arithmetic mean.
paste() - Paste converts its arguments to character strings, and concatenates

them. If the arguments are vectors, they are concatenated term-by-term to give a
character vector result.
ggplot() - initializes a ggplot object. It can be used to declare the input data
frame for a graphic and to specify the set of plot aesthetics intended to be common
throughout all subsequent layers unless specifically overridden.
The dataset was analyzed to get details such as the file size, number of records,
the fields specified and their meaning.
Dataset Size No. of No. of Attribute Names

Name Attributes Records
salaries.csv 707 KB 5 23956 yearID, teamID, lgID, playerID,
salary
master.csv 2.31 MB 24 18355 playerID, birthYear,
birthMonth, birthDay,
birthCountry, birthState,
birthCity, deathYear,
deathMonth, deathDay,
deathCountry, deathState,
deathCity, nameFirst,
nameLast, nameGiven, weight,
height, bats, throws, debut,
finalGame, retroID, bbrefID
batting.csv 6.27 MB 24 97890 playerID, yearID, stint,
teamID, lgID, G, G_batting,
AB, R, H, 2B, 3B, HR, RBI, SB,
CS, BB, SO, IBB, HBP, SH, SF,
GIDP, G_old
US Baseball team datasets salaries.csv, master.csv and batting.csv are loaded

into the R tool and the data field organization and its important dimensions are
understood. It is noted that each record in the master has a corresponding record
in the salaries and batting datasets. The relationships between these datasets were
manually analyzed. The datasets are loaded using the read.csv() function of the R
Package. After this the required packages for this project are installed and loaded
using the below commands.
> install.packages(“data.table”)
> library(data.table)
> install.packages(“Rcpp”, dependencies = TRUE)
> library(Rcpp)
> install.packages(“ggplot2”)
> library(ggplot2)
Data preprocessing describes any type of processing performed on raw data to

prepare it for further analysis. In this step, datasets are processed using data.table
package and their functions. These pre-processing steps are very essential to get
accuracy.
The Pre-Processing steps are:

(i) Filtering
In this process, the salaries dataset is filtered based on their fields in
different dimensions such as filtering records specific to a particular periods
and specific leagues (NL or AL), filtering the records of the players in the
years in which they have not had a chance to bat (AB > 0).
(ii) Sorting
Sorting is the process for ordering the data in a correct format. Here the
filtered data should be sorted based on the chronological order of records
(iii)Summarizing
This deals with summarising data to get minimum, maximum and average
salary year wise, team wise and league wise.
(iv) Merging
This step deal with merging the salary data of the base ball players with
their biographical (personal) data (To combine player’s last and first name
into single name), merging the salary data with batting statistics (history
data) data based on the year, player, Team and League fields (To study
salary given based on performance), merging the salary, master and batting
275 Case Studies
datasets (To study player wise batting performance), filtering the records
of the players in the years in which they have not had a chance to bat (AB
> 0) and merging the three datasets to get a study of player wise batting
performance.
Exploration and Visualization deals with finding the trend of Salaries of Players
over the Years (Entire Data Set), finding trend of Salaries of Players over the Years
>= 1990 and only for the American League, finding the year wise Average Salary,
finding Year wise & League wise Average Salary, finding Year wise & Team wise
Average Salary, finding the correlation between the players Hits and Runs, finding
one Players Salary details in the different years and the teams he belonged to,
finding Batting Average of the Players, visualizing the graph of Players Salary Vs
Height, Players Salary Vs Weight and Players Salary Vs Birth Country. All these
exploration and visualization are done using the function ggplot() that belongs to
the package ggplot2.
The code for the entire case study is as follows.

> setwd(“F:”)
#Loading Data
> salaries<-read.csv(“salaries.csv”,header=TRUE)
> master<-read.csv(“master.csv”,header=TRUE)
> batting<-read.csv(“batting.csv”,header=TRUE)
#Transforming into data table

> salaries=as.data.table(salaries)
> batting=as.data.table(batting)
> master=as.data.table(master)
#Data Filtered, sorted and grouped

> salaries.filtered=salaries[lgID == “AL” & yearID >= 1990, ]
> salaries.filtered.sorted=salaries.filtered[order(salary), ]
> summarized.year=salaries[ ,list(Average=mean(salary)), by=”yearID”]

> summarized.lg=salaries[ ,list(Average=mean(salary),Maximum
=max(salary),minimum=min(salary)),by=”lgID”]
> summarized.year.lg=salaries[,list(Average=mean(salary),Maximum=max(salary),
minimum=min(salary)),by=c(“yearID”,”lgID”)]
> summarized.year.play=salaries[,list(Average=mean(salary),Maximum=max(salary),
minimum=min(salary)),by=c(“yearID”,”playerID”)]
> summarized.year.team=salaries[,list(Average=mean(salary),Maximum=max(salary),
minimum=min(salary)),by=c(“yearID”,”teamID”)]
> batting=as.data.table(batting)
#Data merged and reduced

> merged.sm=merge(salaries,master,by=”playerID”)
> merged.sm$name<-paste(merged.sm$nameFirst,merged.sm$nameLast)
> merged.batting= merge(batting, salaries, by=c(“playerID”, “yearID”, “teamID”,
“lgID”), all.x=TRUE)
> merged.all=merge(merged.batting,master,by=”playerID”)
> merged.all = merged.all[AB > 0, ]
> merged.all[, name:=paste(nameFirst, nameLast)]
> salaries$salary.reduced<-paste(salaries$salary/10000)
> z <- salaries$salary.reduced
> salary.reduced.round.<-round(as.numeric(z,1))
#Visualizing Trend of Salaries

> ggplot(salaries, aes(x=yearID, y=salary.reduced.round.))+geom_point()+
ylab(“Salary (in 10000 USD)”)+xlab(“Year”)
> salaries.filtered$salary.reduced<-paste(salaries.filtered$salary/10000)
> y<-salaries.filtered$salary.reduced
277 Case Studies
> salaries.reduced.round<-round(as.numeric(y,1))
#Visualizing Trend of Salaries of Year > 1990 for American League

> ggplot(salaries.filtered, aes(x=yearID, y=salaries.reduced.round))+geom_point()+
> summarized.year$average.reduced<-paste(summarized.year$Average/10000)
> x<-summarized.year$average.reduced
> average.reduced.round<-round(as.numeric(x,1))
#Visualizing Year wise Average Salary

> ggplot(summarized.year, aes(x=yearID, y=average.reduced.round))+geom_line()+
> summarized.year.lg$average.reduce.round<-paste(summarized.year.
lg$Average/10000)
> averages.reduce.round<-round(as.numeric(summarized.year.lg$average.reduce.
round,1))
#Visualizing Year wise and League wise Average Salary

> ggplot(summarized.year.lg, aes(x=yearID, y=averages.reduce.round, color=lgID))+
geom_smooth(se=FALSE)+ylab (“Average (in 10000 USD)”)+xlab(“Year”)
> summarized.year.team$average.reduce.round<-
paste(summarized.year.team$Average/10000)
> averages.reduce.round <-
round(as.numeric(summarized.year.team$average.reduce.round,1))
#Visualizing Year wise and Team wise Average Salary

> ggplot(summarized.year.team, aes(x=yearID, y=averages.reduce.round,
color=teamID))+geom_path()+ylab(“Average (in 10000 USD)”)+xlab(“Year”)
#Visualizing Correlation between the Players Hits and Runs

> ggplot(merged.all, aes(x=H, y=R,color=Total.HR))+geom_point()
> salaries[salaries$playerID == “aardsda01”]

> merged.all$batting.average<-paste(merged.all$AB/merged.all$H)
#Visualizing Batting Average of Players

> ggplot(merged.all, aes(x=batting.average))+geom_histogram(stat = “count”)
> merged.all$salary.reduced<-paste(merged.all$salary/10000)
> salary.reduced.round<-round(as.numeric(merged.all$salary.reduced,1))
#Visualizing Players Salary Vs. Height

> ggplot(merged.all, aes(x=height.in.inches.,y=salary.reduced.round))+
geom_smooth(se=FALSE)+
ylab(“Salary (in 10000 USD)”)+xlab(“Height (in inches)”)
#Visualizing Players Salary Vs. Weight

> ggplot(merged.all, aes(x=weight.in.pound.,
y=salary.reduced.round,color=weight.in.pound.))+
geom_point()+ylab(“Salary (in 10000 USD)”)+xlab(“Weight (in pound)”)
#Visualizing Players Salary Vs. Birth Country

> ggplot(merged.all, aes(x=birthCountry, y=salary.reduced.
round,color=birthCountry))+
geom_point()+theme(text=element_text(size = 10),
axis.text.x=element_text(angle=90,hjust=1))+
ylab(“Salary (in 10000 USD)”)+xlab(“Birth Country”)
279 Case Studies
Figure 7.21 Ggplot of Trend of Salaries
Figure 7.22 Ggplot of Trend of Salaries of Year > 1990 for American League
Figure 7.23 Ggplot of Year wise Average Salary
Figure 7.24 Ggplot of Year wise and League wise Average Salary
281 Case Studies
Figure 7.25 Ggplot of Year wise and Team wise Average Salary
Figure 7.26 Ggplot of Correlation between the Players Hits and Runs
Figure 7.27 Ggplot of Batting Average of Players
Figure 7.28 Ggplot of Players Salary Vs. Height

283 Case Studies
Figure 7.29 Ggplot of Players Salary Vs. Weight
Figure 7.30 Ggplot of Players Salary Vs. Birth Country
7.5. Social Network Analysis

Social Network Analysis (SNA) deals with finding social structures using the
concepts of networks and graph theory. It characterizes networked structures in
terms of nodes and the ties, edges, or links that connect them. Social Network
Analysis in R is done using the package igraph. The data used for analysis is the
twitter text data used in the Text Mining Section. The terms in this data can be
considered as people and the tweets as groups in the LinkedIn. The term-document
matrix is a representation of the group membership of people.
In this case study, we first build the network of terms based on their co-
occurrence in the same tweets and then build a network of tweets based on the
terms shared by them. After this, we also build a two-mode network composed of
both terms and tweets.
As a first step we build the term-document matrix as in the Text Mining section
using the below code.
> library(tm)
#Reading the input file
> dat<-read.csv(“GameReview.csv”,stringsAsFactors = FALSE)
#Converting it to a corpus
> corp <- Corpus(VectorSource(dat$text))
#Preprocessing – removing stop words, punctuations, whitespaces etc.
> corp <- tm_map(corp, content_transformer(tolower))
> corp <- tm_map(corp, removePunctuation)
> corp <- tm_map(corp, stripWhitespace)
> corp <- tm_map(corp, removeWords, stopwords(“english”))
#Converting into a term-document matrix
> tdm <- TermDocumentMatrix(corp)
#Removing sparse terms
> tdm2 <- removeSparseTerms(tdm, sparse=0.96)
#converting into a matrix
> termDocMatrix <- as.matrix(tdm2)
> termDocMatrix <- termDocMatrix[ ,1:150]
After that the term-document matrix is transformed into a term-term adjacency

matrix, based on which a graph is built. The graph is then plotted to show the
285 Case Studies
relationship between the frequent terms and also make the graph more readable by
setting colors, font sizes and transparency of vertices and edges.
> termDocMatrix[5:10,1:20]
Docs
Terms 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20
good 0 1 0 0 1 1 0 0 0 0 0 0 0 0 0 0 1 1 0 0
great 0 1 0 0 1 0 0 0 1 0 2 0 0 0 1 0 0 0 0 0
money 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
play 0 1 0 0 0 2 0 0 0 0 0 1 0 0 0 0 0 0 0 0
fun 0 0 2 0 0 0 0 0 0 0 0 1 0 0 0 2 0 0 0 0
much 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
> termDocMatrix[termDocMatrix>=1] <- 1
> termMatrix <- termDocMatrix %*% t(termDocMatrix)
> termMatrix[5:10,5:10]
Terms
Terms good great money play fun much
good 29 6 3 8 4 1
great 6 25 1 5 5 1
money 3 1 9 3 1 2
play 8 5 3 25 10 2
fun 4 5 1 10 35 4
much 1 1 2 2 4 8
In the above code, %*% is an operator for the product of two matrices, and the
function t() transposes a matrix. We then build a term-term adjacency matrix. In
this matrix the rows and columns represent terms, and every entry is the number
of concurrences of two terms. Next we can build a graph with the function graph.
adjacency() from package igraph.
> library(igraph)
> g <- graph.adjacency(termMatrix, weighted=T, mode=”undirected”)

> g <- simplify(g)
> V(g)$label <- V(g)$name
> V(g)$degree <- degree(g)
After that, we plot the network with the function layout.fruchterman.reingold()

as in the Fig. 7.31.
> set.seed(3952)
> layout1 <- layout.fruchterman.reingold(g)
> plot(g, layout=layout1)
Figure 7.31 Network of Terms Based on their Co-occurrence
Next, we set the label size of vertices based on their degrees, to make important
terms stand out. Similarly, we also set the width and transparency of edges based
on their weights. This is useful in applications where graphs are crowded with many
vertices and edges. The vertices and edges in the below code are accessed with V()
and E(). The function rgb(red, green, blue, alpha) defines the colors. With the same
layout, we plot the graph again as in Fig. 7.32.
> V(g)$label.cex <- 2.2 * V(g)$degree / max(V(g)$degree)+ .2
> V(g)$label.color <- rgb(0, 0, .2, .8)
287 Case Studies
> V(g)$frame.color <- NA

> egam <- (log(E(g)$weight)+.4) / max(log(E(g)$weight)+.4)
> E(g)$color <- rgb(.5, .5, 0, egam)
> E(g)$width <- egam
Figure 7.32 Network of Terms with Label Size of Vertices Based on their Degrees
Next, we try to detect communities from the graph, called as cohesive blocks
and then plot the network of terms based on cohesive blocks as in Fig. 7.33.
> blocks <- cohesive.blocks(g)
> blocks
Cohesive block structure:
B-1 c 15, n 31
‘- B-2 c 16, n 30 oooooooooo .ooooooooo oooooooooo o
‘- B-3 c 17, n 28 ooo.oooooo .ooooooooo oooooooooo .
> plot(blocks, g, vertex.size=.3, vertex.label.cex=1.5, edge.olor=rgb(.4,.4,0,.3))
Figure 7.33 Network of Terms with Communities called as Cohesive Blocks
Figure 7.34 Network of Terms Based on Maximal Cliques
Next we plot the network of terms based on maximal cliques as in Fig. 7.34.
> cl <- maximal.cliques(g)
> length(cl)
[1] 286
> colbar <- rainbow(length(cl) + 1)
289 Case Studies
> for (i in 1:length(cl)) {

+ V(g)[cl[[i]]]$color <- colbar[i+1]
+}
> plot(g, mark.groups=cl, vertex.size=.3, vertex.label.cex=1.5, edge
color=rgb(.4,.4,0,.3))
Next we plot the network of terms based on largest cliques as in Fig. 7.35.
> cl <- largest.cliques(g)
> length(cl)
[1] 41
> colbar <- rainbow(length(cl) + 1)
> for (i in 1:length(cl)) {
+ V(g)[cl[[i]]]$color <- colbar[i+1]
+}
> plot(g, mark.groups=cl, vertex.size=.3, vertex.label.cex=1.5, edge.
color=rgb(.4,.4,0,.3))
Figure 7.35 Network of Terms Based on Largest Cliques
It is also possible to build a graph of tweets based on the number of terms

that they have in common. Because most tweets contain the words “game”, “play”,
most tweets are connected with others and the graph of tweets is very crowded.
To simplify the graph and find relationship between tweets beyond the above two
keywords, we remove the two words before building a graph.
> idx <- which(dimnames(termDocMatrix)$Terms %in% c(“game”, “play”))
> M <- termDocMatrix[-idx,]
> tweetMatrix <- t(M) %*% M
> g <- graph.adjacency(tweetMatrix, weighted=T, mode = “undirected”)
> g <- simplify(g)
> V(g)$label.cex <- 1
> V(g)$label.color <- rgb(.4, 0, 0, .7)
> V(g)$size <- 2
Next, we have a look at the distribution of degree of vertices and the result is
shown in the below bar graph as in Fig. 7.36. We can see that there are around 20
isolated vertices (with a degree of zero). Note that most of them are caused by the
removal of the two keywords, “game” and “play”.
> barplot(table(V(g)$degree))
Figure 7.36 Distribution of Degree

291 Case Studies
With the code below, the vertex colours are set based on degree, and labels
of isolated vertices are set to tweet IDs and the first 10 characters of every tweet.
The labels of other vertices are set to tweet IDs only, so that the graph will not
be overcrowded with labels. The colour and width of edges are set based on their
weights. The produced graph is shown in Fig. 7.37.
> idx <- V(g)$degree == 0
> V(g)$label.color[idx] <- rgb(0, 0, .3, .7)
> V(g)$label[idx] <- paste(V(g)$name[idx], substr(dat$text[idx], 1, 10), sep=”: “)
> egam <- (log(E(g)$weight)+.2) / max(log(E(g)$weight)+.2)
> E(g)$color <- rgb(.5, .5, 0, egam)
> E(g)$width <- egam
> set.seed(3152)
> layout2 <- layout.fruchterman.reingold(g)
Figure 7.37 Network of Tweets - I
The vertices in crescent are isolated from all others, and next they are removed
from the graph with the function delete.vertices() and re-plot the graph as in Fig. 7.38.
> g2 <- delete.vertices(g, V(g)[degree(g)==0])
> plot(g2, layout=layout.fruchterman.reingold)
Figure 7.38 Network of Tweets - II
Similarly, it is also possible to remove the edges with low degrees to simplify
the graph using the function delete.edges(). After removing edges, some vertices
become isolated and they are also removed. The produced graph is as in Fig. 7.39.
> g3 <- delete.edges(g, E(g)[E(g)$weight <= 1])
> g3 <- delete.vertices(g3, V(g3)[degree(g3) == 0])
Figure 7.39 Network of Tweets - III

293 Case Studies
In the Fig. 7.39, there are some groups (or cliques) of tweets. Few of them are
listed below. The group of tweets (25, 35, 112) is about the word “Awesome”, the
group of tweets (31, 47, 122) is about the word “good” and the group of tweets (57,
58, 67, 75, 103, 146) is about the word “addictive”.
> dat$text[c(25,35,112)]
[1] “ Awesome! A lot of fun!!”
[2] “ Awesome Mysterious Game!! Fun game to play @ night before bed to wind
down!!”
[3] “ Miss Awesome fun”
> dat$text[c(31,47,122)]
[1] “ Error in patching Every time I try to log it it says error in patching but overall good
game.”
[2] “ Good For spending time while waiting for an appointment”
[3] “ Good It is a good game to play while wasting time”
> dat$text[c(57,58,67,75,103,146)]
[1] “ Addictive fun Perfect fun”
[2] “ Wonderful Is a great game and addictive. Brilliant”
[3] “ Addictive Great looking, fun game”
[4] “ ADDICTIVE!!!! This is a fun and easy to play and lose!!”
[5] “ Very fun Addictive game, similar to a Tomogotchi. You will want to check in on your
village and clan. Building, building, building and re-arranging you village. Some battles too.
Ver well constructed.”
[6] “ JD Very addictive fun gaming”
Next we build a two-mode network, which is composed of two types of

vertices, the tweets and the terms. At first, we generate a graph g directly from the
termDocMatrix. After that, different colors and sizes are assigned to term vertices
and tweet vertices. Then, we set the width and color of the edges. The graph is then
plotted with the function layout.fruchterman.reingold() as in Fig. 7.40.
> g <- graph.incidence(termDocMatrix, mode=c(“all”))
> nTerms <- nrow(M)
> nDocs <- ncol(M)

> idx.terms <- 1:nTerms
> idx.docs <- (nTerms+1):(nTerms+nDocs)
> V(g)$color[idx.terms] <- rgb(0, 1, 0, .5)
> V(g)$size[idx.terms] <- 6
> V(g)$color[idx.docs] <- rgb(1, 0, 0, .4)
> V(g)$size[idx.docs] <- 4
> V(g)$label.color <- rgb(0, 0, 0, 0.5)
> V(g)$label.cex <- 1.4*V(g)$degree/max(V(g)$degree) + 1
> E(g)$width <- .3
> E(g)$color <- rgb(.5, .5, 0, .3)
> set.seed(958)
> plot(g, layout=layout.fruchterman.reingold)
Figure 7.40 Two-Mode Network of Terms and Tweets - I

295 Case Studies
The Fig. 7.40 shows that most tweets are around two centers, “game” and “play”.
Next, let’s have a look at which tweets are about “game”. In the code below, the
function nei() returns all vertices which are neighbors of the vertex “game”.
> V(g)[nei(“game”)]
+ 89/181 vertices, named:
[1] 2 3 4 5 6 8 9 11 13 15 17 18 20 21 27 28 29
[18] 30 31 34 35 37 38 39 40 42 44 51 53 54 55 58 59 60
[35] 61 63 64 66 67 71 72 73 76 80 81 82 83 85 87 90 91
[52] 92 93 94 95 97 98 99 101 102 103 105 107 108 109 110 111 115
[69] 116 117 118 119 120 122 125 127 128 129 131 134 136 138 140 141 143
[86] 144 145 148 149
An alternative way is using the function neighborhood() as below.

> V(g)[neighborhood(g, order=1, “game”)[[1]]]
[1] game 2 3 4 5 6 8 9 11 13 15 17 18
[14] 20 21 27 28 29 30 31 34 35 37 38 39 40
[27] 42 44 51 53 54 55 58 59 60 61 63 64 66
[40] 67 71 72 73 76 80 81 82 83 85 87 90 91
[53] 92 93 94 95 97 98 99 101 102 103 105 107 108
[66] 109 110 111 115 116 117 118 119 120 122 125 127 128
[79] 129 131 134 136 138 140 141 143 144 145 148 149
We can also have a further look at which tweets contain all two terms: “game”
and “play”.
> (rdmVertices <- V(g)[nei(“game”) & nei(“play”)])
[1] 2 6 34 35 37 42 44 59 61 66 73 82 92 107 122 131 134
[18] 143 144 149
> dat$text[as.numeric(rdmVertices$label)]
[1] “ Great game I love this game. Unlike other games they constantly give you money to
play. They are always given you a bone. Keep up the good work.”
[2] “ Meh Used to be good until World Cup upgrade.\nNow it lags all the time, making
it difficult to play.\nMaybe if you spent more time getting the game to actually work and less
time trying to squeeze advertising into every nook of game play, we could have a winner.”
...
Next, we remove “game” and “play” to show the relationship between tweets
with other words. Isolated vertices are also deleted from graph.
> idx <- which(V(g)$name %in% c(“game”, “play”))
> g2 <- delete.vertices(g, V(g)[idx-1])
> g2 <- delete.vertices(g2, V(g2)[degree(g2)==0])
> set.seed(209)
From Fig. 7.41, we can clearly see groups of tweets and their keywords, such as
“addictive”, “good” and “fun”.
Figure 7.41 – Two-Mode Network of Terms and Tweets - II

297 Case Studies
 HIGHLIGHTS
 Text mining involves the process of preprocessing the input text, deriving
patterns within the preprocessed data, and finally evaluation of the
output.
 A word cloud is used to present important words in documents.
 Corpus is a collection of text documents.
 The most accurate and highly used credit scoring measure is the Probability
of Default called the PD.
 The function importance() displays the features importance using the
“mean decrease accuracy” measure.
 Common metrics calculated from the confusion matrix are Precision,
Accuracy, True Positive Rate (TP Rate) and False Positive Rate (FP Rate).
 The US crime dataset is used for the EDA of crimes in US
 Pre-processing done in this case study are removing duplicate records,
records with missing values, records with incorrect values, formatting
Timestamp field, binning time intervals (4 intervals – each 6 hours) and
grouping of similar crimes.
 The objective of the EDA on baseball data is to identify trend of base ball
players salary over the years, to understand correlation between players
salary and their performances, analyze if age, country, height and weight
of the players have impact on their performance.
 Social Network Analysis (SNA) is the process of investigating social
structures through the use of networks and graph theory.
 The package used for Text Mining is tm and the package for Social Network
Analysis is igraph.
 The function nei() returns all vertices which are neighbours of the given
vertex.
Glossary
ANCOVA Analysis of Covariance (ANCOVA) is used to test the

main and interaction effects of categorical variables
on a continuous dependent variable, controlling for
the effects of selected other continuous variables,
which co-vary with the dependent.
ANOVA Analysis of variance (ANOVA) is a collection of

statistical models and their procedures which are
used to observe differences between the means of
three or more variables in a population based on the
sample presented.
APRIORI The Apriori Algorithm is an influential algorithm

for mining frequent itemsets for boolean association
rules. Apriori uses a “bottom up” approach, where
frequent subsets are extended one item at a time,
and groups of candidates are tested against the data.
Association Rule Association rule mining is a procedure which is

Mining meant to find frequent patterns, correlations,
associations, or causal structures from data sets
found in various kinds of databases such as relational
databases, transactional databases, and other forms
of data repositories.
Base Environment The functions and the variables from the R’s base
package are stored in the base environment.
Basic Data Types The basic data types in R are Numeric, Integer,
Complex, Logical and Character.
Binomial Binomial Distribution is a frequency distribution

Distribution of the possible number of successful outcomes in a
given number of trials in each of which there is the
same probability of success.
Chi Square Test Chi-square test is a statistical method assessing the

goodness of fit between a set of observed values and
those expected theoretically.
Classification Classification is a data mining function that assigns

items in a collection to target categories or classes.
The goal of classification is to accurately predict the
target class for each case in the data.
Closure In R, a function and its environment together is

called a closure.
Clustering Cluster analysis or clustering is the task of grouping

a set of objects in such a way that objects in the
same group (called a cluster) are more similar to
each other than to those in other groups (clusters).
Confidence Confidence is the percentage of cases containing A

that also contain B where A and B are two disjoint
item sets. Confidence(A =>
B) = P(B|A) = P(A U B) / P(A)
Confusion matrix A confusion matrix is a table that is often used to

describe the performance of a classification model
(or “classifier”) on a set of test data.
Correlation Analysis Correlation analysis is a method of statistical

evaluation used to study the strength of a
relationship between two, numerically measured,
continuous variables (e.g. height and weight).
CRAN CRAN is a network of ftp and web servers around

the world that store identical, up-to-date, versions
of code and documentation for R.
301 Glossary
Data Frame A data frame is a table or a two-dimensional array-

like structure in which each column contains values
of one variable and each row contains one set of
values from each column.
Decision Tree A decision tree is a decision support tool that uses

a tree-like graph or model of decisions and their
possible consequences, including chance event
outcomes, resource costs, and utility. It is one way
to display a classification algorithm.
Density Based Clustering based on density (local cluster criterion),

Clustering such as density-connected points or based on an
explicitly constructed density function.
Dimensionality Dimensionality reduction is the process of

Reduction reducing the number of random variables under
consideration, via obtaining a set of principal
variables. It can be divided into feature selection
and feature extraction.
Document-term A document-term matrix or term-document

matrix matrix is a mathematical matrix that describes
the frequency of terms that occur in a collection
of documents. In a document-term matrix, rows
correspond to documents in the collection and
columns correspond to terms.
Environment In R the variables that we create need to be stored

in an environment. Environments are another type
of variables. We can assign them, manipulate them
and pass them as arguments to functions.
Exploratory Data Exploratory Data Analysis (EDA) is an approach

Analysis to analyzing data sets to summarize their main
characteristics, often with visual methods.
Factors Factors in R are stored as a vector of integer values

with a corresponding set of character values to use
when the factor is displayed. The factor function is
used to create a factor.
Feature Selection Feature selection, also known as variable selection,

attribute selection or variable subset selection, is
the process of selecting a subset of relevant features
(variables, predictors) for use in model construction.
F-value The F critical value is what is referred to as the F

statistic. In general, if your calculated F statistic in
a test is larger than your table F value, you can reject
the null hypothesis.
GitHub GitHub is a web-based Git or version control

repository and Internet hosting service. It is mostly
used for code.
Global Environment An environment is just a place to store variables.

If we start up R and make an assignment, we are
adding an entry in the global environment.
Hierarchical In data mining, hierarchical clustering (also called

Clustering hierarchical cluster analysis or HCA) is a method of
cluster analysis which seeks to build a hierarchy of
clusters.
Histogram Histogram is a diagram consisting of rectangles

whose area is proportional to the frequency of
a variable and whose width is equal to the class
interval.
Home Directory The R home directory is the top-level directory of

the R installation being run. The R home directory
is often referred to as R_HOME.
303 Glossary
Hypothesis Testing Hypothesis Testing is the theory, methods, and

practice of testing a hypothesis by comparing it
with the null hypothesis. The null hypothesis is only
rejected if its probability falls below a predetermined
significance level, in which case the hypothesis
being tested is said to have that level of significance.
Imputation Imputation is the process of replacing missing data

with substituted values.
JSON JSON (JavaScript Object Notation) is a lightweight

data-interchange format.
K-Means Clustering K-Means clustering aims to partition n observations

into k clusters in which each observation belongs
to the cluster with the nearest mean, serving as a
prototype of the cluster. This results in a partitioning
of the data space into Voronoi cells.
K-Medoids Clustering The K-Medoids Clustering algorithm is a clustering

algorithm related to the K-Means algorithm.
K-Medoids chooses data points as centres and works
with a generalization of the Manhattan Norm to
define distance between data points.
Least Square Linear Least Squares Regression is the most widely

Regression used modeling method in which the sum of the
squares of the vertical distances of different points
from the regression curve is minimized.
Linear Regression Linear Regression is an approach for modeling the

relationship between a scalar dependent variable
y and one or more explanatory variables (or
independent variables) denoted X.
Local Outlier Factor The local outlier factor (LOF) is an algorithm for
finding anomalous data points by measuring the
local deviation of a given data point with respect to
its neighbours.
Logistic Regression Logistic regression, is a regression model where the

dependent variable (DV) is categorical.
Lower Tail Test In Hypothesis Testing, the alternative hypothesis

can take the form - H1: μ < μ0 , where a decrease is
hypothesized and this is called a lower-tailed test.
Mean Decrease Mean Decrease Accuracy is a measure for feature

Accuracy importance in Random Forests. This is assessed for
each feature by removing the association between
that feature and the target.
Multiple Regression Multiple regression is an extension of simple linear

regression. It is used when we want to predict the
value of a variable based on the value of two or more
other variables. The variable we want to predict is
called the dependent variable.
Multivariate Outlier A multivariate outlier is a combination of unusual

scores on at least two variables.
Normal Distribution The normal (or Gaussian) distribution is a very

common continuous probability distribution.
Normal distributions are important in statistics and
are often used in the natural and social sciences
to represent real-valued random variables whose
distributions are not known.
Null Hypothesis A null hypothesis is a type of hypothesis used

in statistics that proposes that no statistical
significance exists in a set of given observations. The
null hypothesis attempts to show that no variation
exists between variables or that a single variable is
no different than its mean.
305 Glossary
One-tailed Test In Hypothesis testing, a one-tailed test is appropriate

if the estimated value may depart from the reference
value in only one direction, for example, whether a
machine produces more than one-percent defective
products.
Outlier Detection In data mining, outlier detection is the identification

of items, events or observations which do not
conform to an expected pattern or other items in
a dataset.
Package A Package is a collection of R functions and datasets.

Currently, the CRAN package repository features
10756 available packages.
Parent Environment All environments are nested and so every

environment has a parent environment. The empty
environment sits at the top of the hierarchy without
any parent.
Poisson Regression Poisson regression assumes the response variable

Y has a Poisson distribution, and assumes the
logarithm of its expected value can be modeled by a
linear combination of unknown parameters.
Principal Component Principal component analysis (PCA) is a statistical

Analysis procedure that uses an orthogonal transformation to
convert a set of observations of possibly correlated
variables into a set of values of linearly uncorrelated
variables called principal components.
Probability of Default Probability of default (PD) is a financial term

describing the likelihood of a default over a
particular time horizon. It provides an estimate
of the likelihood that a borrower will be unable to
meet its debt obligations.
P-value When you perform a hypothesis test in statistics,

a p-value helps you determine the significance of
your results. The p-value is a number between 0
and 1 and if the p-value (typically ≤ 0.05) indicates
strong evidence against the null hypothesis, so we
can reject the null hypothesis.
Quantile Quantiles are cutpoints dividing the range of a

probability distribution into contiguous intervals
with equal probabilities, or dividing the observations
in a sample in the same way.
Regression Analysis Regression Analysis is a statistical process for

estimating the relationships among variables.
It includes many techniques for modeling and
analyzing several variables, when the focus is on the
relationship between a dependent variable and one
or more independent variables (or ‘predictors’).
R-Forge R-Forge strives to provide a collaborative

environment for R package developers.
Scatter Plot A scatter plot is a type of plot using Cartesian

coordinates to display values for typically two
variables for a set of data. The data is displayed as
a collection of points, each having the value of one
variable determining the position on the horizontal
axis and the value of the other variable determining
the position on the vertical axis.
Significance Level The significance level, also denoted as alpha or α, is

the probability of rejecting the null hypothesis when
it is true. For example, a significance level of 0.05
indicates a 5% risk of concluding that a difference
exists when there is no actual difference.
307 Glossary
Silhouette Silhouette analysis can be used to study the separation

distance between the resulting clusters. The silhouette
plot displays a measure of how close each point in one
cluster is to points in the neighboring clusters and
thus provides a way to assess parameters like number
of clusters visually.
SMOTE The SMOTE function handles unbalanced

classification problems and it can generate a new
“SMOTEd” data set that addresses the class unbalance
problem.
Social Network Social Network Analysis (SNA) is the process of

Analysis investigating social structures through the use of
networks and graph theory.
Stemming Stemming is the process of reducing inflected (or

sometimes derived) words to their word stem, base or
root form, generally a written word form.
Stop-word A Stop Word is a commonly used word (such as “the”)

that a search engine has been programmed to ignore,
both when indexing entries for searching and when
retrieving them as the result of a search query.
Support Support is the percentage of cases in the data that

contains both A and B, where A and B are two disjoint
item sets. Support(A => B) = P(A U B)
Term-document A document-term matrix or term-document matrix is

Matrix a mathematical matrix that describes the frequency
of terms that occur in a collection of documents.
In a document-term matrix, rows correspond to
documents in the collection and columns correspond
to terms.
Term-term Adjacency A Term-term Adjacency Matrix is a square matrix used

Matrix to represent a finite graph in which the elements of
the matrix indicate whether pairs of terms (vertices)
are adjacent or not in the graph.
Text Mining Text mining, is the process of deriving high-quality

information from text, which is derived through the
devising of patterns and trends such as statistical
pattern learning or data mining.
Transpose The transpose of a matrix is an operator which flips

a matrix over its diagonal, that is it switches the
row and column indices of the matrix by producing
another matrix.
Two-tailed test A two-tailed test is a statistical test in which the

critical area of a distribution is two-sided and tests
whether a sample is greater than or less than a certain
range of values. If the sample being tested falls into
either of the critical areas, the alternative hypothesis
is accepted instead of the null hypothesis.
Type I error In hypothesis testing, a type I error is the incorrect

rejection of a true null hypothesis (a “false positive”).
Type II error In hypothesis testing, a type II error is incorrectly

retaining a false null hypothesis (a “false negative”).
Univariate Outlier A univariate outlier is a data point that consists

of an extreme value on one variable. Outliers can
influence the outcome of statistical analyses.
Upper Tail Test In Hypothesis Testing, the alternative hypothesis

can take the form - H1: μ > μ 0 , where μ0 is
the comparator or null value and an increase is
hypothesized - this type of test is called an upper-
tailed test.
Wordcloud A word cloud is a visual representation of text data,

typically used to depict keyword on websites, or to
visualize free form text in which the importance of
each word is shown with font size or color.
Packages Used
Package Description Page
arules Package for Mining Association Rules and 113, 131, 151
Frequent Itemsets
arulesViz Package for Visualizing Association Rules and 113, 134, 151
Frequent Itemsets
base The R’s base package 7, 10, 56, 72
caret Package for Classification and Regression 113, 122,

Training 147, 151
chron Package for Chronological Objects which can 8, 43, 170,

Handle Dates and Times. 171, 172
cluster Package for “Finding Groups in Data” Cluster 113, 114,

Analysis Extended 117, 151, 160
corrplot Package for Visualizing Correlation Matrix 97, 113, 147
data.table Package which is an extension of data.frame 180, 181
dbConnect Package that Provides a GUI to connect with 55

databases that use MySQL
DBI Package for R Database Interface 55, 70
DMwR Package that provide Functions and Data for 113, 139,
“Data Mining with R” 151, 160, 163
ellipse Package that contains Functions for drawing 160, 165

ellipses and ellipse-like confidence regions
foreign Package for Reading and Writing Data Stored 54

by Minitab, S, SAS, SPSS, Stata, Systat,
Weka, dBase, etc.
fpc Package that contains Flexible procedures for 113, 114,

Clustering 116, 118,
119, 120
ggplot2 Package that Creates Elegant Data 155, 170,

Visulaizations Using the Grammar of 172, 180, 181
Graphics
graphics The R Graphics Package 7
grDevices The R Graphics Devices and Support for 7

Colours and Fonts
igraph Package for Network Analysis and 190

Visualization
lattice Package for Trellis Graphics in R 81, 82, 83,

85, 86, 87
lubridate Package that Makes Dealing with Dates a 46, 47, 49

Little Easier
MASS Package that Support Functions and 61, 62, 109,

Datasets for Veneble’s and Ripley’s MASS 110
methods Package for formal methods and classes 7

311 Packages Used
party A Package for Recursive Partytioning 113, 122,

123, 151
plotrix Package that contains Various Plotting 73, 86, 87

Functions
plyr Package with Tools for Splitting, Applying 58, 60, 113,
and Combining Data 147, 151
randomForest Package for Random Forests for 130, 149,

Classification and Regression 151, 160, 165
Rcpp Package for R and C++ Integration 171, 181
reshape2 Package that Flexibly Reshapes Data 63
rpart Package for Recursive Partitioning and 127, 130,

Regression Trees 151, 161, 167
RSQLite Package that provides SQLite interface for 55

R
sqldf Package that Performs SQL Selects on R Data 52, 60, 61, 70
Frames
stats The R Package for Statistics 112, 113,

114, 136, 151
stringr Simple, Consistent Wrappers for Common 56

String Operations
TH.data Package for TH’s Data Archive 125

Tm Text Mining Package 152, 153
utils The R Utils Package 7
wordcloud Package that draws Word Cloud 152, 156
xlsx Package that provides Read, Write and 54

Format of Excel Files
XML Package that provide Tools for Parsing and 53

Generating XML within R
Functions Used
Function Package Description Page
.libPaths() base .libPaths gets/sets the library trees 7

within which packages are looked for.
abline() graphics This function adds one or more 100, 102,

straight lines through the current 126
plot.
acos() base This function computes the arc- 6

cosine.
aes() ggplot2 Aesthetic mappings describe how 86, 155,

variables in the data are mapped 174, 182,
to visual properties (aesthetics) of 183
geoms.
aggregate() stats Splits the data into subsets, computes 63, 70,
summary statistics for each, and 171,172,
returns the result in a convenient 174,
form.
all.equal() base all.equal(x, y) is a utility to 6

compare R objects x and y testing
‘near equality’.
anova() stats Compute analysis of variance (or 107, 108,

deviance) tables for one or more 109, 112
fitted model objects.
aov() stats Fit an analysis of variance model by a 107,108,

call to lm for each stratum. 109, 112,
apply() base Returns a vector or array or list 63, 70,

of values obtained by applying a 145
function to margins of an array or
matrix.
apriori() arules This function Mines frequent 131,

itemsets, association rules or 134,151,
association hyperedges using the
Apriori algorithm.
args() base Displays the argument names and 10.11

corresponding default values of a
function or primitive.
arrange() plyr This function completes the 60,149

subsetting, transforming and
ordering triad with a function
that works in a similar way to
subset and transform but for
reordering a data frame by its
columns.
array() base Creates or tests for arrays. 23, 64
as.character() base Create or test for objects of 19,30,172,

type “character”. 173
as.data.frame() base Functions to check if an object 130

is a data frame, or coerce it if
possible.
315 Functions Used
as.data.table() data. Functions to check if an object 180, 182

table is data.table, or coerce it if
possible.
as.Date() base Functions to convert 43, 45, 49,

between character 58
representations and objects
of class “Date” representing
calendar dates.
as.environment base A generic function coercing 9

() an R object to an environment.
as.integer() base Creates or tests for objects of 18

type “integer”.
as.list() base Functions to construct, coerce 9, 29, 30

and check for both kinds
of R lists.
as.numeric() base Creates or coerces objects of 30, 165, 182,

type “numeric”. is.numeric is a 183
more general test of an object
being interpretable as numbers.
as.POSIXlt() base Functions to manipulate 42, 43, 44,

objects of classes 170, 172
“POSIXlt” and
“POSIXct”
representing calendar
dates and times.
as.yaml() yaml Convert an R object into a 53

YAML string
asin() base This function computes the arc- 6

sine.
assign() base Assign a value to a name in an 4, 9,16

environment.
atan() base This function computes the 6

arc-tan.
available. utils Returns a matrix of 8

packages() details corresponding
to packages currently
available at one or more
repositories.
BarChart() lattice Plots a bar chart, such as 85,87

for counts, with default
colors for one or two
variables.
barplot() graphics Creates a bar plot with 85, 87, 155

vertical or horizontal
bars.
basename() base Removes all of the path 42

up to and including the
last path separator.
biplot() stats Plot a biplot on the 140, 151

current graphics device.
body() base Get or set the body of a 10, 11

function.
317 Functions Used
boxplot() graphics Produce box-and- 87, 137,

whisker plot(s) of the 162, 163
given (grouped) values.
boxplot. grDevices This function is typically 136, 137,

stats() called by another 151
function to gather the
statistics necessary for
producing box plots,
but may be invoked
separately.
brewer.pal() RColorBrewer Creates nice looking 156

color palettes especially
for thematic maps.
bwplot() lattice Produces box-and- 82,87

whisker plots
by() base Function by is an 63,67, 68,

object-oriented wrapper 70,
for tapply applied to
data frames.
c() base This is a generic function 5, 16, 22,

which combines its 30, 39
arguments.
cat() base Outputs the objects, 39,40, 159

concatenating the
representations. cat
performs much less
conversion thanprint.
cbind() base Take a sequence of 25, 49, 61,

vector, matrix or data- 130
frame arguments and
combine by columns
or rows, respectively.
ceiling_date() lubridate Takes a date-time object 48

and rounds it up to the
nearest boundary of the
specified time unit.
chisq.test() stats Performs chi-squared contingency 109, 110,

table tests and goodness-of-fit tests. 112
clara() cluster Computes a “clara” object, a list 116, 151

representing a clustering of the
data into k clusters.
class() base A simple generic function 18,43

mechanism which can be used
for an object-oriented style of
programming.
cmdscale() stats Classical multidimensional 164

scaling (MDS) of a data
matrix. Also known as principal
coordinates analysis.
coef() stats A generic function which extracts 101

model coefficients from objects
returned by modeling functions.
319 Functions Used
cohesive. igraph Calculates cohesive blocks for 191

blocks() objects of class igraph.
colMeans() base Form row and column sums and 35, 36,
means for numeric arrays. 147
colnames() base Retrieve or set the row or column 24, 31,

names of a matrix-like object. 59, 73,
88
colSums() base Form row and column sums and 35, 36,
means for numeric arrays. 133,
complete. stats Return a logical vector indicating 59

cases() which cases are complete, i.e.,
have no missing values.
confint() stats Computes confidence intervals 106

for one or more parameters in a
fitted model.
content_ tm Create content transformers, 153

transformer() i.e., functions which modify the
content of an R object.
coord_flip() ggplot2 Flip cartesian coordinates so that 86, 155

horizontal becomes vertical, and
vertical, horizontal.
cor() stats Computes the correlation 91, 96,

of x and y if these are vectors. 112, 165
cor.test() stats Test for association between paired 96, 97,

samples, using one of Pearson’s 112
product moment correlation
coefficient, Kendall’s tau or
Spearman’s rho.
Corpus() tm Representing and computing on 153

corpora.
corrplot() corrplot A graphical display of a correlation 97, 112,

matrix, confidence interval. 148,
cov() stats Compute the covariance 91

of x and y if these are vectors.
ctree() party Recursive partitioning for 122, 123,

continuous, censored, ordered, 124, 130,
nominal and multivariate response 151
variables in a conditional inference
framework.
cummax() base Returns a vector whose elements 91

are the cumulative maxima of the
elements of the argument.
cummin() base Returns a vector whose elements 91

are the cumulative minima of the
cumprod() base Returns a vector whose elements 91

are the cumulativeproducts of the
cumsum() base Returns a vector whose elements 91

are the cumulative sums of the
321 Functions Used
cut() base Divides the range of x into intervals 38, 171,

and codes the values in x according 172
to which interval they fall.
cutree() stats Cuts a tree, as resulting from 118, 119,

hclust, into several groups either by 157
specifying the desired number(s) of
groups or the cut height(s).
daisy() cluster Compute all the pairwise 114, 160,

dissimilarities (distances) between 163, 164,
observations in the data set.
data() utils Loads specified data sets, 50, 74

or list the available data
sets.
data.frame() base This function creates data 11, 32, 58,

frames, tightly coupled 60, 61,
collections of variables 100, 130,
which share many of the 137, 149,
properties of matrices 155, 167,
and of lists, used as 168
the fundamental data
structure by most of R’s
modeling software.
days() chron Returns Various Periods 47, 49

from a Chron or Dates
Object
dbConnect() dbConnect Creates a connection to a 55

DBMS
dbDisconnect() DBI Function to disconnect 55

from a database
dbDriver() DBI Function to create a new 55

driver object given the
name of a database
dbGetQuery() DBI Function to Send query, 55

retrieve results and then
clear result set. Returns
the result of a query as a
data frame.
dbinom() stats Density, distribution 95, 112

function, quantile
function and random
generation for the
binomial distribution with
parameters size and prob.
dbListTables() DBI This function returns 55

a character vector that
enumerates all tables and
views in the database.
dbReadTable() DBI This Function returns a 55

data frame that contains
the complete data from
the remote table.
dbscan() fpc Generates a density based 114, 119,

clustering of arbitrary 120, 141,
shape. 151
dbUnloadDriver() DBI Function to remove a 55

driver object given the
name of a database
323 Functions Used
dcast() reshape2 Cast functions Cast a

molten data frame into an
array or data frame.
ddays() lubridate Quickly creates Duration 46, 47

days for easy date-time
manipulation.
degree() igraph The degree of a vertex is its most 190, 194,

basic structural property, the 195, 196,
number of its adjacent edges. 197, 198
delete.edges() igraph Deletes edges from a graph. 196
delete. igraph Deletes vertices from a graph. 195, 196,

vertices() 198
density() stats The function density computes 139

kernel density estimates. Its
default method does so with the
given kernel and bandwidth for
univariate observations.
deparse() base Turn unevaluated expressions 10, 11

into character strings.
dhours() lubridate Quickly creates Duration hours 46, 49

for easy date-time manipulation.
difftime() base Time intervals creation, printing, 45

and some arithmetic.
dim() base Retrieve or set the dimension of 24, 27,

an object. 118, 144,
147, 148
dimnames() base Retrieve or set the dimnames of 24, 31,

an object. 154,
dirname() base basename removes all of the 42

path up to and including the last
path separator.
dist() stats This function computes and 118, 157,

returns the distance matrix
computed by using the specified
distance measure to compute
the distances between the rows
of a data matrix.
dminutes() lubridate Quickly creates Duration 46, 49

minutes for easy date-time
manipulation.
dmy() lubridate Parse dates according to the 46

order in that day, month, and
year elements appear in the
input vector.Transforms dates
stored in character and numeric
vectors to Date or POSIXct
objects.
325 Functions Used
dnorm() stats Density, distribution function, 92, 93

quantile function and random
generation for the normal
distribution with mean equal
to mean and standard deviation
equal to sd.
do.call() base do.call constructs and executes a 11

function call from a name or a
function and a list of arguments
to be passed to it.
download.file() utils This function can be used 54

to download a file from the
Internet.
droplevels() base The function droplevels is used 37

to drop unused levels from a
factor or, more commonly, from
factors in a data frame.
dseconds() lubridate Quickly creates Duration 46, 49

seconds for easy date-time
manipulation.
dweeks() lubridate Quickly creates Duration weeks 46, 47, 49

dyears() lubridate Quickly creates Duration years 46, 47, 49

dym() lubridate Parse dates according to the 46

order in that day, year, and
month elements appear in the
input vector.
eclat() arules This funciton Mines frequent 131, 151

itemsets with the Eclat
algorithm.
exists() base Look for an R object of the given 9, 10

name and possibly return it
facet_wrap() ggplot2 This function wraps a 1d 77, 87

sequence of panels into 2d. This
is generally a better use of screen
space than facet_grid because
most displays are roughly
rectangular.
factor() base The function factor is used 36, 66,

to encode a vector as a factor 139, 149,
(the terms ‘category’ and 151, 163,
‘enumerated type’ are also used 164, 171
for factors).
file.path() base Construct the path to a file 42

from components in a platform-
independent way.
findAssocs() tm Find associations in a document- 156

term or term-document matrix.
findCorrelation() caret This function searches through a 148

correlation matrix and returns a
vector of integers corresponding
to columns to remove to reduce
pair-wise correlations.
327 Functions Used
findFreqTerms() tm Find frequent terms in a 155

document-term or term-
document matrix.
fixed() stringr Compare literal bytes in the 56, 57

string. This is very fast, but not
usually what you want for non-
ASCII character sets.
floor_date() lubridate This function takes a date-time 48

object and rounds it down to the
nearest boundary of the specified
time unit.
force_tz() lubridate This function returns a the date- 48

time that has the same clock
time as x in the new time zone.
formalArgs() methods These are utilities, currently in 10,11

the methods package, that either
provide some functionality
needed by the package.
formals() base Get or set the formal arguments 10

of a function.
format() base Format an R object for pretty 40, 46,

printing. 171
formatC() base Formatting numbers individually 40

and flexibly, formatC() using
C style format specifications.
fromJSON() jsonlite This function is used to convert 53

between JSON data and R
objects.
geom_bar() ggplot2 This function makes the height 86, 87,

of the bar proportional to the 155,
number of cases in each group.
grep() base Search for matches to 56, 70,

argument patternwithin each 154
element of a character vector: they
differ in the format of and amount
of detail in the results.
grepl() base Search for matches to 56

argument patternwithin each
gsub() base Search for matches to 56, 70

argument patternwithin each
hclust() stats Hierarchical cluster analysis on a 114, 118,

set of dissimilarities and methods 119, 151,
for analyzing it. 157,
head() utils Returns the first or last parts of a 32, 49,

vector, matrix, table, data frame or 50, 52,
function. Since head() andtail() are 62, 137,
generic functions, they may also
have been extended to other classes.
329 Functions Used
help() utils help is the primary interface to the 4, 16

help systems.
help. utils Allows for searching the help 4

search() system for documentation
matching a given character string
in the (file) name, alias, title,
concept or keyword entries (or
any combination thereof), using
either fuzzy matching or regular
expression matching.
hist() graphics The generic function hist computes 80, 81,

a histogram of the given data values. 87, 95
If plot = TRUE, the resulting
object of class “histogram” is
plotted by plot.histogram, before it
is returned.
histogram() lattice Draw Histograms, possibly 183

conditioned on other variables.
hours() chron Return Hours from a Times Object. 47, 49
htmlParse() XML This function Parses a HTML file 53

or HTML content, and generates
an R structure representing the
HTML tree.
htmlTreeParse() XML Parses an HTML file or string 53

containing HTML content,
and generates an R structure
representing the HTML tree.
identical() base The safe and reliable way to test two 5, 60

objects for being exactly equal.
It returns TRUE in this
case,FALSE in every other case.
ifelse() base Returns a value with the same 13, 16,

shape as test which is filled 164, 171,
with elements selected from 172, 173,
either yes or no depending
on whether the element
of test is TRUE or FALSE.
importance() random This is the extractor function for 129, 151,

Forest variable importance measures as 165, 166
produced by randomForest.
inspect() arules Provides the generic function 131, 132,

inspect to display associations 133, 134,
plus additional information 154
formatted for online inspection.
install. utils Download and install packages 8, 16, 46,

packages() from CRAN-like repositories or 51, 73,
from local files. 171, 181
installed. utils Find (or retrieve) details of 7

packages() all packages installed in the
specified libraries.
interaction() base Computes a factor which 39, 49

represents the interaction
of the given factors. The
result of interaction is always
unordered.
331 Functions Used
intersect() base Performs the set operation 138

intersect on the two member
vectors
is.atomic() base is.atomic returns TRUE if x is 27

of an atomic type (or NULL)
and FALSE otherwise.
is.integer() base Creates or tests for objects of 18

type “integer”.
is.na() base NA is a logical constant of length 37,

1 which contains a missing value 147,
indicator. NA can be coerced to 162,
any other vector type except raw. 170,
172
is.recursive() base is.atomic returns TRUE if x is 27

of an atomic type (or NULL)
and FALSE otherwise.
isTRUE() base These operators act on raw, logical 6

and number-like vectors.
kmeans() stats Perform k-means clustering on a 114,

data matrix. 115,
143,
151,
158
knnImputation() DMwR Function that fills in all NA values 160

using the k Nearest Neighbours of
each case with NA values.
lapply() base lapply returns a list of the same 63,

length as X, each element of which 64,
is the result of applying FUNto the 65,
corresponding element of X. 153,
largest.cliques() igraph This function finds all, the largest 193

cliques in an undirected graph.
The size of the largest clique can
also be calculated.
layout() base layout divides the device up into 74,

as many rows and columns as 87,
there are in matrix mat, with 117,
the column-widths and the row-
heights specified in the respective
arguments.
layout. igraph This is a deprecated layout 190,

fruchterman. function. 194,
reingold() 197
legend() graphics This function can be used to add 72,

legends to plots. Note that a call to 79,
the function locator(1) can be used 84, 85
in place of the x and y arguments.
length() base Get or set the length of vectors 21,

(including lists) and factors, and 24,
of any other R object for which a 31,
method has been defined. 56,
65,
72,
107,
173
333 Functions Used
match() base match returns a vector of the 90, 112

positions of (first) matches of its
first argument in its second.
matrix() base matrix creates a matrix from the 23, 25,

given set of values. 63, 74,
117, 121
max() base Returns the (parallel) maxima and 88, 91,

minima of the input values. 112, 156,
182
maximal. igraph This function finds all the maximal 193

cliques() cliques in an undirected graph.
mdy() lubridate Parse dates according to the order in 46

that month, day, and year elements
appear in the input vector.
mean() base Generic function for the (trimmed) 5, 16, 45,

arithmetic mean. 46, 88,
89, 180,
182
median() stats Compute the sample median. 5, 16, 88,

89
melt() reshape2 Convert an object into a molten 62, 63, 70

data frame.
merge() base Merge two data frames by common 35, 49,

columns or row names, or do other 62, 180,
versions of database joinoperations. 182
message() base Generate a diagnostic message from 12, 13,

its arguments. 14, 15
min() base Returns the (parallel) maxima and 88,91,

minima of the input values. 112, 162,
180, 182
minutes() chron Return Minutes from a Times 47

Object.
months() base This function Extracts the month. 171, 172
mutate() plyr Mutate a data frame by adding new 58

or replacing existing columns.
myd() lubridate Parse dates according to the order in 46

that month, year, and day elements
appear in the input vector.
na.fail() stats Returns the object if it does not 59

contain any missing values, and
signals an error otherwise.
na.omit() stats Returns the object with 59, 70

incomplete cases removed.
names() base Functions to get or set the 2, 26, 31,

names of an object. 131, 155,
174
ncol() base Returns the number of columns 24, 27, 31

present in x.
335 Functions Used
nearZeroVar() caret This function diagnoses 147

predictors that have one
unique value or predictors that
have both of the following
characteristics: they have very
few unique values relative to
the number of samples and
the ratio of the frequency of
the most common value to the
frequency of the second most
common value is large.
nei() igraph This function finds the vertices 198

not farther than a given limit
from another fixed vertex, these
are called the neighborhood of
the vertex.
neighborhood() igraph This function finds the vertices 198

not farther than a given limit
from another fixed vertex, these
are called the neighborhood of
the vertex.
new.env() base Get, set, test for and create 8, 9, 16

environments.
nlevels() base Return the number of levels 36

which its argument has.
nls() stats Determine the nonlinear 105, 106

(weighted) least-squares
estimates of the parameters of
a nonlinear model.
noquote() base Print character strings without 40

quotes.
nrow() base Returns the number of rows 24, 59,

present in x. 128, 140,
162
odbcClose() RODBC Close connections to ODBC 55

databases.
odbcConnect() RODBC Open connections to ODBC 55

databases.
order() base Returns a permutation which 61,70,

rearranges its first argument 143, 180,
into ascending or descending
order, breaking ties by further
arguments. sort.list is the same,
using only one argument.
outliers. DMwR This function uses hierarchical 160, 163

ranking() clustering to obtain a ranking of
outliers for a set of cases.
pairs() graphics A matrix of scatterplots is 75, 87,

produced. 140
pam() cluster Partitioning (clustering) of the 114, 116,

data into k clusters “around 117, 118,
medoids”, a more robust version 151,
of K-means.
pamk() fpc Partitioning around medoids 114, 116,

with estimation of number of 117
clusters
337 Functions Used
par() graphics Used to set or query graphical 84, 85

parameters. Parameters can
be set by specifying them as
arguments to par.
parent.frame() base These functions provide access 10

to environments (‘frames’ in S
terminology) associated with
functions further up the calling
stack.
paste() base Concatenate vectors after 19, 39,

converting to character. 49, 166,
182, 183
paste0() base Concatenate vectors after 39

converting to character.
path. base Expand a path name, for 7

expand(“~”) example by replacing a leading
tilde by the user’s home
directory (if defined on that
platform).
pbinom() stats Density, distribution function, 95, 96,

quantile function and 112
random generation for the
pie() graphics Draw a pie chart. 72, 87
pie3D() plotrix Displays a 3D pie chart with 73, 87

optional labels.
plot() graphics Generic function for plotting 73, 100,

of R objects. For more details 126, 142,
about the graphical parameter 157, 164,
arguments, see par. 166,
plotcluster() fpc Plots to distinguish given classes 120

by ten available projection
methods. One-dimensional data
is plotted against the cluster
number.
plotcorr() ellipse This function plots a correlation 160, 165

matrix using ellipse-shaped
glyphs for each entry. The
ellipse represents a level curve of
the density of a bivariate normal
with the matching correlation.
pmax() base Returns the (parallel) maxima 91

and minima of the input values.
pmin() base Returns the (parallel) maxima 91

and minima of the input values.
pnorm() stats Density, distribution function, 92, 93,

quantile function and random 112
generation for the normal
distribution with mean equal
to mean and standard deviation
equal to sd.
prcomp() stats Performs a principal components 140, 145,

analysis on the given data matrix 151
and returns the results as an
object of class prcomp.
339 Functions Used
predict() stats predict is a generic function 99, 100,

for predictions from the 122, 123,
results of various model fitting 124, 125,
functions. The function invokes 126, 128,
particular methods which 129, 168
depend on the class of the first
argument.
print() base print prints its 4, 61,

argument and returns 124, 126,
it invisibly (via invisible(x)). It is 127, 139,
a generic function which means 143, 149
that new printing methods can
be easily added for new classes.
printcp() rpart Displays the cp table for fitted 167

rpart object.
Density, distribution function,

quantile function and
95, 96,
rbinom() stats random generation for the
112
52, 54,
Reads a file in table format and
59, 162,
creates a data frame from it, with
read.csv() utils 170, 171,
cases corresponding to lines and
172, 180,
variables to fields in the file.
182

read.csv2() utils 52

read.delim() utils 52

read.delim2() utils 52
This function reads a file stored

read.spss() foreign by the SPSS save or export 54, 70
commands.
Generates a SAS program to

convert the ssd contents to SAS
read.ssd() foreign 54, 70
transport format and then uses
read.xport to obtain a data frame.

read.table() utils 51. 52, 54
Read the contents of a worksheet

read.xlsx() xlsx 53, 70
into an R data.frame.
Read the contents of a worksheet

read.xlsx2() xlsx 53, 54, 70
into an R data.frame.
Read some or all text lines from

readLines() base 10, 52, 70
a connection.
341 Functions Used
readMat() R.matlab Reads matlab files 54, 70
Draws rectangles around the

branches of a dendrogram
highlighting the corresponding
118, 119,
rect.hclust() stats clusters. First the dendrogram
157
is cut at a certain level, then
a rectangle is drawn around
selected branches.
The levels of a factor are

re-ordered so that the level
specified by ref is first and
the others are moved down.
relevel() stats 37
This is useful for contr.
treatment contrasts which
take the first level as the
reference.
Removes installed packages/

remove.packages() utils bundles and updates index 8
information as necessary.
Remove sparse terms from

remove Sparse
tm a document-term or term- 82
Terms()
document matrix.
reorder is a generic function.

The “default” method
treats its first argument as
reorder() stats a categorical variable, and 82, 83
reorders its levels based on the
values of a second variable,
usually numeric.
rep replicates the values in x.

22, 23,
It is a generic function, and
rep() base 130, 140,
the (internal) default method
146,
is described here.
lapply returns a list of the

same length as X, each
replicate() base element of which is the 12
result of applying FUN to the
corresponding element of X.
library and require load and

require() base 7, 163
attach add-on packages.
This function reshapes a data

frame between ‘wide’ format
with repeated measurements
in separate columns of
reshape() stats
the same record and ‘long’
format with the repeated
measurements in separate
records.
residuals is a generic function

which extracts model residuals
resid() stats 106
from objects returned by
modeling functions.
343 Functions Used
This function shows the

cross-validated prediction
performance of models
with sequentially reduced
rfcv() randomForest 160, 166
number of predictors (ranked
by variable importance) via
a nested cross-validation
procedure.
This function creates colors

corresponding to the given 191, 192,
rgb() grDevices intensities (between 0 and 193, 194,
max) of the red, green and 197
blue primaries.
Density, distribution function,

quantile function and random 12, 92,
generation for the normal 94, 95,
rnorm() stats
distribution with mean 112, 136,
equal to mean and standard 137
deviation equal to sd.
This function rounds the

131, 132,
values in its first argument
round() base 164, 182,
to the specified number of
183
decimal places (default 0).
Form row and column sums

rowMeans() base 35, 36
and means for numeric arrays.
Retrieve or set the row or

24, 31,
rownames() base column names of a matrix-like
145, 149
object.
Form row and column sums 35, 36,

rowSums() base
and means for numeric arrays. 143, 155
122, 125,
Recursive Partitioning and
127, 130,
rpart() rpart Regression Trees. Fit a rpart
151, 161,
model.
167
This function provide

information about the
uniform distribution on the
runif() stats 121
interval from min to max.
The function runif generates
random deviates.
sample takes a sample of

38, 123,
the specified size from the
sample() base 127, 128,
elements of x using either with
164
or without replacement.
sapply() base Returns a list of the same 63, 65,

length as X, each element 70, 148
of which is the result
of applying FUN to the
scan() base Read data into a vector or list 52

from the console or file.
sd() stats This function computes 90, 112

the standard deviation
of the values in x. If na.
rm is TRUE then missing
values are removed before
computation proceeds.
345 Functions Used
search() base Gives a list 7

of attached packages,
and R objects, usually data.
frames.
seconds() chron Return Seconds from a Times 47, 49

Object.
seq() base Generate regular 20, 45,

sequences. seq is a standard 46, 64,
generic with a default 92, 93,
method. 94, 95,
106
seq.int() base Generate regular 20, 38

sequences. seq.int is a
primitive which can be
much faster but has a few
restrictions.
seq_along() base Generate regular 20

sequences. seq_
along and seq_len are very
fast primitives for two
common cases.
seq_len() base Generate regular 20

sequences. seq_
along and seq_len are very
fast primitives for two
common cases.
set.seed() base This function is the 121, 125,

recommended way to specify 127, 128,
seeds for Random Number 136, 147,
Generation. 156, 158,
165
setRepositories() utils Interact with the user 8

to choose the package
repositories to be used.
setwd() base Used to set the working 41, 42,

directory to dir. 49, 172,
182
simplify() igraph Creates simple graphs. 190, 194

Simple graphs are graphs
which do not contain loop
and multiple edges.
SMOTE() DMwR This function handles 160, 164

unbalanced classification
problems using the SMOTE
method. Namely, it can
generate a new “SMOTEd”
data set that addresses the
class unbalance problem.
solve() base This generic function solves 26

the equation a %*% x = b
for x, where b can be either a
vector or a matrix.
347 Functions Used
sort() base Sort (or order) a vector 59, 60,

or factor (partially) into 70, 132,
ascending or descending 134, 156,
order. For ordering along 159
more than one variable,
e.g., for sorting data frames,
see order.
sqldf() sqldf SQL select on data frames 61
sqlQuery() RODBC Submit an SQL query to an 55

ODBC database, and retrieve
the results.
sqrt() base This function computes the 6, 143

square root of x.
stamp() lubridate Stamps are just like format, 46

but based on human-frendly
templates like “Recorded at
10 am, September 2002” or
“Meeting, Sunday May 1,
2000, at 10:20 pm”.
stemCompletion() tm Heuristically complete 153

stemmed words.
stopwords() tm Return various kinds of 153

stopwords with support for
different languages.
str() utils Compactly display the 51, 54,

internal structure of 144
an R object, a diagnostic
function and an alternative
to summary (and to some
extent, dput).
str_count() stringr Count the number of 57, 70

matches in a string.
str_detect() stringr Detect the presence or 56, 57

absence of a pattern in a
string.
str_replace() stringr Replace matched patterns in 57, 70

a string.
str_replace_all() stringr Replace matched patterns in 57

a string.
str_split() stringr Split up a string into pieces. 57, 70
str_split_fixed() stringr Split up a string into pieces. 57
strftime() base Functions to convert 44

between character
representations and objects
of classes “POSIXlt”
and ”POSIXct”
representing calendar
dates and times.
stripWhitespace() tm Strip extra whitespace from 153

a text document. Multiple
whitespace characters are
collapsed to a single blank.
349 Functions Used
strptime() base Functions to convert 43, 44,

between character 172,
representations and
objects of classes “POSIXlt”
and ”POSIXct” representing
calendar dates and times.
strsplit() base Split the elements of a 41

character vector
x into substrings according
to the
matches to
substringsplit within them.
structure() base This function returns the

given object with further
attributes set.
strwrap() base Each character string in the 153

input is
first split into paragraphs.
sub() base Search for matches to 20,70

argument
pattern within each element
of a character vector.
subset() base Return subsets of vectors, 34, 155,

matrices or 170, 172
data frames which meet
conditions.
substr() base Extract or replace substrings 20, 41

in a character vector.
substring() base Extract or replace substrings 41

in a character vector.
sum() base sum returns the sum of all 3, 4, 5

the values present in its
arguments.
summary() base summary is a generic function 45, 46,

used to produce result 67, 91,
summaries of the results 99, 103,
of various model fitting 104,
functions. The function invokes 1112,
particular methods which 131, 136,
depend on the class of the first 148,
argument.
switch() base switch evaluates EXPR and 13, 14,

accordingly chooses one of the 16
further arguments.
Sys.getenv base Sys.getenv obtains the values of 7

(“HOME”) the environment variables.
Sys.getlocale() base Get details of or set aspects of 44

the locale for the R process.
Sys.time() base Sys.time and Sys.Date returns 42, 43,

the system’s idea of the current 44, 48,
date with and without time. 49
Sys.timezone() base Information about time zones 44, 49

in R. Sys.timezone returns the
name of the current time zone.
351 Functions Used
system.file() base Finds the full file names of files 51, 53,
in packages etc. 55, 70,
t() base Given a matrix or data. 25, 34,

frame x, t returns the transpose
of x.
table() base table uses the cross-classifying 38, 114,

factors to build a contingency 127, 151,
table of the counts at each 168, 171
combination of factor levels.
tabulate() base tabulate takes the integer- 90, 112

valued vector bin and counts the
number of times each integer
occurs in it.
tan() base These functions give the obvious 6

trigonometric functions. They
respectively compute the cosine,
sine, tangent, arc-cosine, arc-
sine, arc-tangent, and the two-
argument arc-tangent.
tapply() base Apply a function to each cell of 63, 66,

a ragged array, that is to each 67, 68,
(non-empty) group of values 69
given by a unique combination
of the levels of certain factors.
TermDocument tm Constructs or coerces to a 154

Matrix() term-document matrix or a
document-term matrix.
text() graphics This function Adds Text to a 125, 127,

Plot 165, 168
theme() ggplot2 Modify components of a 77, 173,

theme 174
times() chron Generate Times 171

Components from Input
tm_map() tm Interface to apply 153

transformation functions to
corpora.
today() lubridate Returns the current date 46, 47, 48
toJSON() jsonlite Converts R objects to JSON 53, 70
tolower() base Translate characters in 41

character vectors, in
particular from upper to
lower case or vice versa.
toString() base This is a helper function 39

for format to produce a single
character string describing
an R object.
toupper() base Translate characters in 41

character vectors, in
particular from upper to
lower case or vice versa.
353 Functions Used
union() base Performs the set operation 138

union on the two member
vectors
unique() base unique returns a vector, data 37, 90,

frame or array like x but with 112, 171,
duplicate elements/rows 173
removed.
unlist() base Given a list 30, 153

structure x, unlist simplifies
it to produce a vector which
contains all the atomic
components which occur
in x.
update() stats update will update and (by 77

default) re-fit a model. It
does this by extracting the
call stored in the object,
updating the call and (by
default) evaluating that call.
update. utils old.packages indicates 8

packages() packages which have a
(suitable) later version on the
repositories whereas update.
Packages offers to download
and install such packages.
vapply() base Returns a list of the same 63, 65,

length as X, each element 70
of which is the result
of applying FUN to the
var() stats Computes the variance of x. 90, 112
varImpPlot() random Dotchart of variable 129, 149,

Forest importance as measured by a 151, 166
Random Forest.
vector() base vector produces a vector of the 20, 49

given length and mode.
VectorSource() tm This function creates a vector 153

source.
View() utils Invoke a spreadsheet-style 7, 8

data viewer on a matrix-
like R object.
weekdays() base This function Extracts the 171, 172

weekday.
weeks() lubridate Quickly create Period object 47, 49

(week) for easy date-time
manipulation
which() base Give the TRUE indices of a 22, 49,

logical object, allowing for 137, 148,
array indices. 170
which.max() base Determines the location, i.e., 22, 90,

index of the (first) minimum 112, 166
or maximum of a numeric (or
logical) vector.
355 Functions Used
which.min() base Determines the location, i.e., 22

index of the (first) minimum
or maximum of a numeric (or
logical) vector.
with() base Evaluate an R expression in an 58

environment constructed from
data, possibly modifying (a
copy of) the original data.
with_tz() lubridate This function returns a date- 48

time as it would appear in a
different time zone.
within() base Evaluate an R expression in 58

an environment constructed
from data, possibly modifying
(a copy of) the original data.
wordcloud() wordcloud Plot a word cloud 156
write.csv() utils Similar to write.table() 52, 70
write.foreign() foreign This function exports 54

simple data frames to other
statistical packages by writing
the data as free-format text
and writing a separate file
of instructions for the other
package to read the data.
write.table() utils write.table prints its required 52, 70

argument x (after converting
it to a data frame if it is not
one nor a matrix) to a file
or connection.
write.xlsx2() xlsx Write a data.frame to an 53, 70

Excel workbook.
writeLines() base Write text lines to a 52, 53,

connection. 70, 153
writeMat() R.matlab Writes matlab files 54, 70
xmlParse() XML This function Parses a XML 53, 70

file or XML content, and
generates an R structure
representing the XML tree.
xmlTreeParse() XML Parses an XML file or string 53

containing XML content,
and generates an R structure
representing the XML tree.
xyplot() lattice This function produces 75, 76,

bivariate scatterplots or 79, 87
time-series plots
yaml.load() yaml Parse a YAML string and 53

return R objects.
yaml.load_ yaml This function calls yaml. 53

file() load with the contents of the
specified file or connection.
357 Functions Used
years() lubridate Quickly create Period object 47, 49,

(year) for easy date-time
manipulation
ymd_hm() lubridate Parse dates that have hours 46

or minutes elements.
ymd_hms() lubridate Parse dates that have 46

hours, minutes, or seconds
elements.
References
BOOKS
1. Paul Teetor, “R Cookbook”, O’Reilly Media, 2011.

2. Ohri, “R for Business Analytics”, Springer, 2012.
3. Yanchang Zhao, “R and Data Mining: Examples and Case Studies”, ELSEVIER,
2012.
4. Richard Cotton, “Learning R - A Step-by-Step Function Guide to Data Analysis”,
O’Reilly Media, 2013.
5. Roger D. Peng, “R Programming for Data Science”, Lean Publishing, 2014.
6. Yanchang Zhao, Yonghua Cen, “Data Mining Applications with R”, ELSEVIER,
2014.
WEBSITES
1. https://www.tutorialspoint.com/r/
2. http://www.r-tutor.com/
3. https://cran.r-project.org/manuals.html
4. https://www.r-bloggers.com/
5. http://www.rdatamining.com/examples/
Index
Symbols Basic Data Types 27, 28, 80, 299

Binomial Distribution 141, 151, 322,
.Library command 11 337, 339
Biplot 215, 216, 224, 231, 316
Box Plot 132, 133, 140, 211, 244, 247
A
Break Statement 22, 26
Byrow 38, 122
Accessing Databases 90
Aggregate 104, 105, 114, 116, 261, 262,
265, 266, 313 C
ANCOVA 141, 170, 176, 299
ANOVA 141, 168, 169, 170, 176, 299 Calculations with Dates and Times 74
APRIORI 202, 231, 299 Vast 83, 103, 104, 116
Arrange 99, 228, 314 Cbind 40, 41, 57, 81, 100, 101, 202,
Array 37, 38, 39, 81, 105, 108, 114, 144, 229, 318
176, 222, 260, 261, 301, 314, 323, Character 10, 15, 24, 31, 32, 35, 49, 50,
351, 353, 354 60, 61, 65, 66, 68, 69, 81, 86, 87,
Association Rule Mining 177, 202, 203, 89, 109, 121, 144, 202, 235, 260-
205, 237 264, 273, 302, 314, 315, 322-324,
327-329, 335, 337, 348, 349, 350,
352
B
Chi Square Test 172, 300
Chron 12, 72, 259, 262, 263, 309, 321,
Balancing 244, 250, 251
329, 334, 345, 352
Bar Plot 238, 316
Clara 179, 182, 230, 318
Base Environment 15, 299
362 Index
Classification 168, 177, 190, 191, 200, Dataset 4, 84-86, 103, 120, 124, 141,
228, 237, 243, 244, 250, 256, 300, 142, 147, 155, 160, 168, 170, 172,
301, 307, 346 173, 176, 179, 182, 188, 190, 191,
Closure 15, 300 194, 197, 199, 202, 211, 220, 221,
Clustering 177-180, 182, 185-188, 217, 222, 224, 230, 233, 243, 244, 247,
218, 230, 233, 237, 240, 242, 244, 249, 250, 253, 255, 257, 258, 261,
248, 300, 302, 303, 318, 322, 331, 262, 273, 274, 297, 305
336 Date Conversion 263
Coef 160, 211, 318 Dates and Times 71, 74, 309
Colbycol 86 Dbinom 151, 152, 176, 322
ColClasses 88, 89 Dbscan 178, 186, 187, 217, 230, 322
Collapse 65, 66, 235 Decision Tree 190, 192, 193, 301
Complex 2, 27, 29, 30, 168, 248, 253 Dendrogram 185, 241, 341
Confidence 201, 202, 300 Density Based Clustering 178, 186-188,
Confusion matrix 300 218, 219, 301
Corpus 234, 235, 284, 297, 320 Density Plot 214
Correlation Analysis xi, 141 Dimensionality Reduction 178, 220,
Corrplot 155, 176, 178, 226, 230, 309, 301
320 DMwR 178, 214, 230, 231, 243, 244,
CRAN 3, 10, 11, 12, 25, 300, 305, 330 248, 249, 250, 309, 331, 336, 346
Credit Scoring 243, 258, 297 Dnorm 148, 149, 176, 325
CSV 86, 87, 89, 90, 97, 115, 226, 233, Document-Term Matrix 233, 236, 301,
234, 247, 260-262, 270-275, 284, 307, 352
339, 355
E
D
Eclat 178, 202, 230, 231, 326
Data Frame 11, 51-56, 60, 81, 84, 85, Environment 10, 13-15, 18, 25, 153,
87, 88, 95-98, 100, 102, 114-116, 160, 299-302, 305, 306, 315, 316,
123, 125, 132, 144, 166, 176, 212, 350, 355
260, 261, 272, 273, 301, 314, 315, Excel 86, 88, 312, 356
322, 323, 325, 328, 333, 334, 339, Exploratory Data Analysis 258
340, 342, 353, 356 Export 83, 88, 115, 340
Data Reshaping x
363 Index
F Hypothesis Testing 141, 173, 308
Factors 28, 60, 81, 161, 302 I

Feature Selection 177, 228, 301
FindfreqTerms 237, 327 If and Else 19
Flow Control 1 Import 83, 86, 88, 89, 115
For Loop 24 Imputations Removal 249
Functions 1, 2, 5, 7, 8, 10, 11, 13, 15- Integer 28, 80, 299
17, 25, 35, 38, 39, 44, 45, 49, 51, Inverse of a Matrix 43
57, 59, 65, 69, 73, 76-80, 82, 83, IQR 145, 176
86-89, 93, 94, 100, 103, 105, 115,
116, 126, 128, 129, 141, 145, 146,
J
148, 151, 153, 166, 175-178, 180,
182, 190, 200, 201, 230, 234, 259,
262, 270, 271, 274, 299, 301, 305, JSON 85, 88, 115, 303, 327, 352
318, 319, 323, 328, 337, 339, 342,
350,-352 K
F-value 168, 302
K-Means Clustering 178, 179, 181, 182,
G 220, 303
K-Medoids Clustering 182, 183, 184,
Gini Gain 228 303
GitHub 12, 302 knnImputation 243, 249, 331
Global Environment 302
L
H
Lattice 118, 123, 125, 126, 128, 130,
Help 1, 3, 4, 5, 9, 10, 25, 329 133, 136, 138, 139, 140, 310, 316,
Hierarchical Clustering 185, 230, 240, 317, 329, 356
244, 248, 302, 336 Least Square Regression 166
Histogram 130, 131, 140, 150, 272, Level 2, 61, 62, 64, 81, 86, 87, 118, 168,
278, 329 173-175, 202, 260, 302, 303, 306,
Home Directory 302 338, 341
HTML 83, 87, 88, 329 LHS 201
364 Index
Lift 201, 202, 203, 205-208 193, 194, 197, 199, 220, 221, 243,
Linear Orthogonal Transformation 221 244, 247, 249, 252, 255, 257, 258,
Linear Regression 156, 159, 303 300, 301, 302, 304, 313, 314, 318,
Line Plot 127 319, 335, 339, 342, 344, 350, 353
List 10, 11, 13, 14, 17, 25, 27, 37, 38, Multiple Regression 161, 304
43-45, 47-51, 64, 68-71, 81, 91, Multivariate Outlier Detection 177
105-109, 114, 124, 125, 166, 167,
205, 208, 210, 211, 214, 222, 227, N
234, 239, 248, 249, 251, 265, 275,
276, 314, 315, 318, 321, 325, 332,
Next Statement 19, 22, 26
336, 342, 344, 345, 353, 354
Normal Distribution 141, 148, 176,
Local Outlier Factor 214, 303
325, 338, 343
Logical 19, 21, 24, 29, 30, 32, 45, 92,
Null Hypothesis 168, 169, 173-175,
119, 132, 144, 162, 176, 260, 319,
302, 303, 304, 306, 308
331, 354, 355
Numeric 28, 29, 32, 35, 48, 49, 59, 64,
Logistic Regression 162, 304
81, 89, 118, 127, 130, 131, 133-
Loop 22, 23, 24, 26, 346
135, 141-145, 147, 153, 157, 158,
Lower Tail Test 173, 304
162, 165, 176, 214, 226, 242-244,
247, 249, 252, 276-278, 295, 315,
M 319, 324, 341, 343, 354, 355
Matlab 89, 115, 341, 356 O

Matrices and Arrays 27, 39, 41
Mean 5, 7, 25, 76, 77, 114, 141-145,
One-tailed Test 305
148-150, 175, 176, 225, 228, 253,
Order 28, 35, 63, 77, 98, 99, 116, 133,
273, 275, 276, 297, 303, 304, 325,
215, 219, 226, 228, 240, 271, 272,
333, 338, 343
274, 275, 295, 324, 325, 333, 334,
Mean Decrease Accuracy 253, 297
336, 347
Median 7, 25, 132, 133, 141, 142, 143,
Outlier Detection xi
144, 175, 333
Outliers.Ranking 244, 249, 336
Merge 58, 81, 100, 102, 272, 276, 333
Outliers Removal 249
MinPts 186, 217
Model 5, 151, 156-160, 162-164, 165-
167, 170-172, 176, 188, 190, 191,
365 Index
P 171, 176, 191, 227

Principal Component Analysis 177,
Package 10, 25, 138, 178, 190, 191, 194, 215, 231
199, 216, 231, 243, 274, 305, 309- Probability of Default 243, 297, 305
314, 316, 318, 320, 321, 323, 325, Prune 206
326, 328, 329, 331, 333, 334, 336, P-value 306
338, 339, 341, 342, 344, 346, 348,
350, 352, 353, 355 Q
Package Party 191
Package RandomForest 199 Qbinom 151, 153, 176
Package rpart 194, 230 Qnorm 148, 150, 176
Packages and Functions for Classifica- Qplot 261, 262, 265
tion 190 Quantile 306
Packages and Functions for Clustering
178
R
Packages for Data Mining 177
Parent Environment 305
Pbinom 151, 152, 176, 337 RandomForest 177, 178, 190, 199, 201,
Pie chart 118, 139, 337 228, 230, 244, 253, 311, 330, 343
Plotcluster 187, 188, 338 Range 71, 130, 142, 145, 176, 196, 244,
Plotcorr 244, 252, 338 247, 249, 260, 306, 308, 321
Plyr 96, 99, 178, 226, 230, 311, 314, 334 Rank 100, 116, 248, 272
Pnorm 148, 149, 176, 338 Ranking Features 253
Poisson Regression 164, 305 Rbind 17, 40, 41, 57, 58, 81, 100, 101
POSIXct 71, 72, 74, 77, 82, 315, 324, Rbinom 151, 153, 176, 339
348, 349 Read.csv 86, 89, 90, 97, 226, 234, 247,
POSIXlt 71, 72-74, 80, 82, 260, 262, 260, 262, 272, 274, 275, 284, 339
263, 315, 348, 349 Regression Analysis 156, 306
Prcomp 215, 216, 223, 231, 338 Reshape 342
Precision 257, 258, 297 Response 156, 157, 159, 160, 162, 164,
Predict 157-159, 161, 167, 189, 190, 165, 166, 170, 176, 305, 320
191, 193, 194, 196-198, 200, 201, rfcv 244, 255, 343
230, 244, 255, 257, 300, 304, 339 R-Forge 11, 12, 306
Predictor 156-160, 162, 164-166, 170, RHS 201
Rlof 178, 216, 230, 231
366 Index
Rnorm 18, 148, 150, 151, 176, 211, 260, 262, 263, 302, 314, 349
212, 343 Summary Statistics 141, 261, 313
RowSums 59, 60, 219, 238, 240, 343 Support 190, 201, 307, 310
RStudio 3 Switch Statement 21
Runif 188, 189, 344
T
S
Term-Document Matrix 236, 237, 239,
SAS 83, 86, 89, 115, 310, 340 242, 284, 301, 307, 326, 327, 341,
Scatter plot 118, 120, 122, 139, 140, 352
187, 212, 213, 218, 306 Term-Term Adjacency Matrix 284, 285
Sequence 4, 6, 32, 33, 76, 318, 326 Text Mining 233, 297, 308
Set.seed 188, 191, 194, 197, 199, 211, Threshold 208, 229, 244, 255
226, 240, 242, 253, 286, 291, 294, Time Zone 73, 74, 77, 80, 82, 327, 350,
296, 346 355
Significance level 173-175, 303, 306 Transpose 40, 56, 242, 308, 351
Silhouette 307 True Positive Rate 257, 297
Skip 22, 26, 86 Tweet 89, 291, 293
SMOTE 244, 250, 307, 346 Two-Tailed Test 308
Social Network Analysis 233 Type I error 308
Sort 98, 99, 116, 205, 208, 228, 240, Type II error 308
242, 336, 347
SPSS 83, 89, 115, 310, 340 U
Sqldf 86, 100, 116, 311, 347
SQLite 90, 116, 311
Univariate Outlier Detection 211, 212
Standard Deviation 141, 145, 148, 176,
Upper Tail Test 174, 308
325, 338, 343, 344
Statistical Analysis 141
Statistical Measures 141 V
Stemming 234, 236
Stop-word 307 Variable Scope 17
Strings 9, 27, 60, 65, 67, 69, 73, 81, 86, Variable Selection 220, 302
88, 93, 98, 116, 273, 323, 335 Variance 145, 168, 170, 176, 221-224,
Subset 56, 62, 114, 160, 206, 238, 253, 241, 299, 313, 314, 354
367 Index
VarImpPlot 200, 229, 230, 253, 254, 176, 182, 183, 185, 191, 193, 194,
354 202, 203, 205-208, 211-214, 216,
Vector 4, 7, 21, 24, 27, 32-37, 43, 47, 227, 234, 239-241, 243, 244, 247,
49, 61, 66, 69, 81, 87, 93, 98, 100, 249, 252, 255, 258-262, 272, 274,
105, 107, 108, 110, 116, 118, 127, 275, 284, 290, 293, 295-297, 299-
130, 132, 134, 135, 141-144, 148, 303, 305, 307-309, 313, 317-319,
151, 157, 158, 166, 176, 180, 222, 321, 330-332, 335, 336, 339, 341,
225, 261, 273, 302, 314, 318-320, 342, 344, 345, 346, 349, 350, 353,
322, 324, 325, 326, 328, 331, 333, 354, 355
334, 344, 346, 347, 349, 350, 351, Wordcloud 233, 239, 240, 312, 355
353, 354, 355 Working Directory 70, 81, 346
W X
Which 2, 4, 5, 12, 17, 18, 20, 21, 27, XML 83, 85, 87, 88, 115, 312, 329, 356
34, 35, 41, 61, 64, 67, 70, 81, 83, XPath 87
84, 87, 88, 102, 105, 108, 115, 118,
119, 125, 126, 130, 132, 134, 139, Y
143, 144, 148-150, 152, 156, 157,
162, 164, 165, 168, 170, 172, 173,
YAML 83, 85, 88, 316, 356

R Programming Text Book

Uploaded by

Copyright:

Available Formats

R Programming Text Book

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

R Programming Text Book

Uploaded by

Copyright:

Available Formats

R Programming

– An Approach to Data Analytics

Libros de Estadística-Ciencia de Datos|Statistics-Data Science Books (PDF)

Assistant Professor, Department of Computer Science,

Principal, Government Arts and Science College,

Chennai New Delhi Tirunelveli

Libros de Estadística-Ciencia de Datos|Statistics-Data Science Books (PDF)

ISBN 978-81-8094-408-6 MJP Publishers

Libros de Estadística-Ciencia de Datos|Statistics-Data Science Books (PDF)

As the authors explain in their book, understanding the techniques and

Mr. Sajeev Madhavan

Libros de Estadística-Ciencia de Datos|Statistics-Data Science Books (PDF)

This book on R-Programming starts with simple concepts like programming

My heartfelt appreciations to the authors who have done a wonderful job in

R is a programming language and a free open source software environment for

Chapter 1 introduces the programming language R, briefs on how to install R

Libros de Estadística-Ciencia de Datos|Statistics-Data Science Books (PDF)

Chapter 6 is about exploring the data mining techniques available in R. It

Libros de Estadística-Ciencia de Datos|Statistics-Data Science Books (PDF)

One of the authors (Dr. G. Sudhamathy) thanks the authorities of Avinashilingam

The author expresses her reverential gratitude to Shri. Dr. P. R. Krishnakumar,

A special thanks to Dr. G. Padmavathi, Professor, Department of Computer

Libros de Estadística-Ciencia de Datos|Statistics-Data Science Books (PDF)

Libros de Estadística-Ciencia de Datos|Statistics-Data Science Books (PDF)

Libros de Estadística-Ciencia de Datos|Statistics-Data Science Books (PDF)

On completion of this Chapter you will be able to:

R is a free open source language that has cross platform compatibility. R is a

R can be installed from the link http://www.r-project.org using internet

Once R installation is completed we install R-Studio. For installation of

R Studio is an Integrated Development Environment (IDE) that consists of

Figure 1.1 R-Studio GUI

1.3.1. First Program

1.3.3. Assigning Variables

1.3.4. Basic Mathematical Operations

The division operator is of three types. The ordinary division is represented

[1] TRUE TRUE TRUE

Non-integers cannot be compared using the operator “==” as it produces

> c(“A”, “B”, “C”) < “B”

The function installed.packages() returns a data frame with information about

The function R.home(“library”) retrieves the location on the machine that

When R is upgraded, it is required to reinstall all the packages as different

1.5. Environments and Functions

In R the variables that we create need to be stored in an environment. Environments

The assign() function can also be used to assign variables to an environment.

> ls(envir = newenvironment)

To create user defined functions, it is required to just assign the function as we

Functions can be passed as arguments to other functions and they can be

When using functions as arguments to the do.call() function, it is not necessary

1.5.3. Variable Scope

> x <- function(a1)

1.6. Flow Control

1.6.1. If and Else Statement

1.6.2. Switch Statement

1.7.1. Repeat Loops

1.7.2. While Loops

1.7.3. For Loops

d <- c(TRUE, FALSE)

 The function repeat is similar to the do while statement in other languages.

On completion of this Chapter you will be able to:

2.1. Basic Data Types in R