2024-2025 BSCM.

BI-01
Business Analytics Portfolio Assignment
Case
For over 35 years, SuperBusiness Inc. has been the industry leader in workspace products like
furniture, technology, and traditional office supplies. The company has been operating profitably for a
long time, but recently, the Management Team (MT) is under the impression that business is not going
as well. Unfortunately, the company has not invested much in data analytics, and is therefore not able
to quantify this “impression.” They have hired lucky you to “do something with data.” They have given
you complete freedom (and the accompanying dataset) to analyze the data and give them
recommendations. For them, any insights are welcome. This is also an opportunity for them to
recognize the added value of a data driven approach.
Your assignment is therefore to examine the dataset, formulate a business problem and corresponding
business objective(s), conduct the corresponding analyses and report back to management (see below
for more details). It is important to remember that you can scope the assignment and dataset anyway
that you want. You can focus on a particular geographic area, product category, time period, type of
customer, type of order etc… Just make sure your objective/ data and analysis are coherent. You can
make up the strategy of the organization, the composition of the MT etc. Be creative.

Dataset Description
The dataset consists of 51290 lines. Each line is one product purchased as part of an order (Order ID x
Product ID). This means that an order can consist of multiple lines. Depending on your objective and
your scope, you will need to manipulate the data (data preparation). For example, for the regression
analysis, you will need to create a time series. For CART, you may want to calculate new variables (e.g.
lead time). In the following table is a description of each variable.

Sample
Variable Value Description
Category Office There are three product
Supplies categories: office supplies,
furniture and technology
City Los Angeles This is the city where the order
was placed.
Country United States This is the country where the
store is located.
Customer LS-172304 Each customer has a unique ID.
ID
Customer Lycoris This is the name of the person
Name Saunders who ordered.
Discount 0 This is the percentage of
discount that they received on
this item/order.
Market US This is a classification. There is
some overlap (e.g. EU/EMEA).
Order Date 1/7/11 This is the date of the order.
Order ID CA-2011- This is a unique identifier for the
130813 order.
Order High Order priority can be low,
Priority medium, high or critical.
Products with high priority need
to be shipped within 2 days,

1
 medium within a week, and low
within a month.
Product ID OFF-PA- This is also the SKU number.
10002005
Product Xerox 225
Name
Profit 9.3312 This is the margin (sales - cost)
on that item for that order. It
can be negative!
Quantity 3 This is the number of items
ordered.
Region West Region where the order was
placed
Row ID 36624 This is a unique identifier for the
row (so only for the dataset)
Sales 19 This is the sales amount
Segment Consumer This is the customer segment
Ship Date 1/9/11 This is the day the product was
shipped.
Ship Mode Second Class
Shipping 4.37 The cost of shipping the order
Cost (not all products are shipped
together)
State California
Sub- Paper
Category
Year 2011 Year in which the order was
placed
Market2 North
America
weeknum 2

Instructions, Tips, Tricks etc.

Learning Objectives (from the student manual):

The BI portfolio shows that you are able to:

1. Evaluate the obtained information from key analytical techniques, using the main assessment
models.
2. Interpret the information stemming from selected key analytical techniques and use this
information for making informed decisions and recommendations.

Background / purpose / big idea behind the assignments:

Business Analytics is an upcoming field that will affect how decisions are made within a supply chain.
The BI portfolio assignment mimics how Business Analytics projects are executed in real life.

What will you actually be doing?

2
Based on this case and the accompanying dataset, you will analyze the dataset and prepare a
management report advising management on a particular problem. Given the information provided in
the case, you are free to come up with your own “angle” and corresponding business objective(s). You
can also add all information that you think is relevant (e.g. strategy etc). This is the perspective from
which you will be analyzing the data. The only requirement is that you conduct two analyses. For each
analysis, you will be applying the steps of the CRISP-DM framework. In analysis I you will be showing
that you can make a classification tree and a regression tree (CART). For analysis II you will be
focusing on forecasting methods, using a time series model and a regression model. You will bundle
these two analyses in one portfolio.

Your portfolio will consist of one executive summary (recommended length 1 A4, with a max of 2 A4),
a management report (4 A4 with a max of 6 A4) with 2 analysis reports in the appendices detailing
your thinking and actions for the CRISP-DM steps. You will also submit your R-scripts as separate files.

The case company description is fictitious, but based on reality. The dataset is real.

Executive Summary
Summarize for the busy executive/ manager that will read your report the following points: your
business objective(s) including motivation, your main findings, how you got to those findings and what
the company should do. The order does not matter, as long as your summary is a persuasive argument
for the reader to want to read more.

Management Report
The management report focusses on the business problem/ objective and the why, how, when, who
etc. The management report is really intended for the business (domain experts) that need your
analysis to solve a problem / take advantage of an opportunity. If you think of it in terms of CRISP-DM:
the management report is mostly the business understanding and deployment steps.

Analysis Reports (x2)

3
CRISP-DM Framework

Business Understanding: Given the assignment description (see above) and the dataset provided,
formulate a business objective for your data mining project. Given the company strategy, explain why
this information is timely and relevant for the company. Translate your business goal to a data mining
goal. In other words, what insights are needed to support decision-making for this issue? You will
notice that, in order to do a good job on this step, you most likely already need to execute part of the
data understanding (and maybe even data preparation steps). This information will also be part of your
management report.

Data Understanding: you will first get to know your data. Based on the business problem, do you
need a target variable? Are you trying to predict something (in which case you need a target variable)
or are you trying to better describe an existing situation (in which the target variable is optional)?
What will that target variable be? Do you have all the variables needed (as far as you can tell) to
answer that question or problem? Is the data of sufficient quality? What level of measurement is your
data? What analysis methods do you think you can and should do? Is the data appropriate for this
analysis? Apply some basic descriptive statistics and make a few graphs to give an idea of the data.

Data Preparation: assuming your data is of sufficient quality, what kind of “manipulations” did you
have to do on the data to prepare it for modeling? How did you deal with missing data? Did you drop
some variables? Did you have to create new ones, combine other ones? Depending on the models that
you will be using, you may have to go back and forth between data preparation and modeling.

Modeling: you will run several models to find out which one is best. Don’t forget to explain why you
made the choices that you made. This step is done individually since you each have your own
question/ problem, but you are welcome to help each other.

Evaluation: which model was best? Use relevant performance evaluation to support your
argumentation. What went well and what went less well? Was there data missing (e.g. variables) that
you think would have resulted in a better model? If the company continues, what improvements do
you think are necessary to the data they collect? What data should they monitor to ensure that your
project remains a success? Summarize the main actions/ decisions/ analytics in a word report (max 8
pages).

4
Deployment: your management report needs to summarizie your findings in a clear, professional,
and visual way. This report is meant for non-experts.