Everand Logo

Welcome to Everand!

  • Discover millions of ebooks, audiobooks, and so much more with a free trial

    Only €10,99/month after trial. Cancel anytime.

    Spatial Statistics Illustrated
    Spatial Statistics Illustrated
    Spatial Statistics Illustrated
    Ebook177 pages1 hour

    Spatial Statistics Illustrated

    By Lauren Bennett and Flora Vale

    5/5

    ()

    Read preview

    About this ebook

    Spatial statistics empowers you to go beyond visual analysis to answer questions confidently and make data-driven decisions.

    Thanks to the data and computational power now at our fingertips, data science is in every aspect of our lives. But with so many algorithms and buzzwords floating around, where do you start to solve complex problems or figure out where to go next?

    There has never been a more exciting time to learn about spatial statistics. Spatial statistics uses an aspect of geography that helps you quantify patterns and relationships so that you can feel confident in your analysis.

    Spatial Statistics Illustrated is an introductory book for learning the concepts behind the powerful spatial statistics tools in ArcGIS.

    With approachable explanations and uncomplicated drawings, Spatial Statistics Illustrated gives readers an accessible understanding of some of the most widely used spatial statistics methods, including how they work and when to use them. In a friendly, conversational tone, the authors share techniques that can help you explore your data in meaningful ways; quantify patterns and relationships; understand trends, and make informed, impactful decisions.

    This book has something for everyone analyzing data, including:

    • seasoned data scientists looking to explore the value that spatial analysis offers
    • GIS analysts looking to expand their spatial statistics skill set
    • new GIS users discovering the value of spatial statistics

    Spatial Statistics Illustrated is a perfect complement to more traditional, technical statistics and spatial statistics texts and is also ideal as supplemental reading for academic courses.

    Based on the popular series of Spatial Statistics workshops presented by the authors at the annual Esri User Conference, Spatial Statistics Illustrated welcomes readers into the unparalleled world of spatial statistics.


    LanguageEnglish
    PublisherEsri Press
    Release dateJun 6, 2023
    ISBN9781589485716
    Spatial Statistics Illustrated
    Author

    Lauren Bennett

    Dr. Lauren Bennett leads the Spatial Analysis and Data Science product engineering team at Esri. In her role, she oversees the R&D of the ArcGIS geoprocessing framework, which includes spatial and spatiotemporal statistics, raster and multidimensional analysis, machine learning and big data analytics. Lauren is passionate about the power of spatial data science to make an impact on our world, from racial equity and social justice to public health to climate change. Lauren received a BA in Geography from McGill University, an MS in Geographic and Cartographic Science from George Mason University, and her PhD in Information Systems and Technology from Claremont Graduate University.

    Related to Spatial Statistics Illustrated

    Related ebooks

    Technology & Engineering For You

    View More
    View More

    Related podcast episodes

    Related categories

    Reviews for Spatial Statistics Illustrated

    5/5

    1 rating0 reviews

    What did you think?

    Tap to rate

    Review must be at least 10 words

      Book preview

      Spatial Statistics Illustrated - Lauren Bennett

      Introduction

      Welcome

      There has never been a more exciting time to learn about analysis. Now, more than ever, our colleagues, communities, and leaders understand the importance of making data-driven decisions. The field of data science, with its focus on using algorithms to turn data into information and knowledge, has brought analysis into just about every aspect of our lives. This explosion has happened in part because of our unprecedented access to data and computational power. With so many algorithms and buzzwords floating around, it can be overwhelming to figure out where to start or where to go next. In this book, we want to offer an approachable introduction to a key area of data science that offers huge opportunities when solving the complex problems our world faces: spatial statistics.

      Spatial statistics is a vast discipline with countless methods and applications. This book focuses specifically on the spatial statistics and spatial machine learning tools and capabilities available in ArcGIS®. We will do our best to explain and illustrate how each method works in a conceptual way. The point isn’t to teach you how to write the algorithms, but rather how to know when to use them, how to apply them, and how to interpret them. We don’t all have to be engineers of machine learning algorithms, but we do have to be knowledgeable about how the methods work so that we can use them appropriately. Our goal is to help you learn about some of the fundamental concepts and amazing algorithms that power this kind of spatial, analytical approach.

      While we’re cautioning that we have to be careful about how we use these methods, we also want to say it’s not rocket science and you really are capable of doing it. Whether you’re new to analysis, or you’ve been doing complex analysis for years, we will try to introduce these methods in a way that will feel approachable so that you can use them to solve important problems.

      We’ll start by exploring what spatial statistics are and the key role that spatial relationships play in how we understand the world around us in chapter 1, "Why spatial is special. Then we’ll look at how some of the most common descriptive statistics are adapted and expanded to summarize the spatial characteristics of our data in chapter 2, Means and medians. Next, we’ll turn to cluster analysis, including both statistical and machine learning–based approaches. First, we’ll explore a few techniques for automating the detection of natural clusters in our data in chapter 3, Finding clusters with machine learning. Then, we’ll discuss some of the most widely used statistical clustering techniques and how they can be applied to both spatial and spatiotemporal data in chapter 4, Statistical cluster analysis, and chapter 5, Spatiotemporal pattern mining. Finally, we’ll introduce some of the foundational concepts in regression analysis and show how our models can be improved by incorporating spatial relationships in chapter 6, Modeling spatial relationships and making predictions."

      Welcome to the beautiful world of spatial statistics.

      Asking the right question

      There are countless methods at our fingertips to solve the complex problems that we face. Whether we use traditional statistical approaches, newer machine learning methods, or dive into deep learning and the broader world of artificial intelligence, the first and arguably most important step in an analysis is figuring out what question we’re trying to answer. It is easy to fall into the trap of I have this data, what analysis should I do? or I have this new tool in the toolbox, what data can I stuff into it? This will rarely lead us to the answers we seek or the solutions to problems we need to solve.

      Instead, we must first and foremost think about what problem we’re trying to solve, and what questions we need to ask to help solve that problem. Let’s take the example of ensuring our city is providing an equitable, high-quality, and accessible public transportation system. There is no single tool or algorithm that will solve this problem.

      It’s a big problem with a lot of complexities and nuance. But we can break it down into a series of answerable questions:

      Who is currently using public transportation?

      Where are the communities with the highest need for public transportation?

      Where are there gaps in access?

      Who is most impacted by those gaps?

      What are the current usage patterns for each type of public transit?

      Are there areas where high usage may indicate the need for additional stops or lines?

      Are there areas that are more likely to experience delays or interruptions in service?

      What is the current distribution of resources, and how does that compare to need?

      Suddenly, that big, complex problem has at least some building-block questions that we can answer. Now that we’ve got these smaller, digestible questions, we can start to figure out which approach will help us answer a particular question. And sometimes that algorithm will be state of the art or cutting edge. But sometimes it will be a method that has been around for decades. We don’t solve our problem more effectively just because we use a more complicated algorithm or a buzzier approach.

      This shift from focusing on the methods to focusing on the question is critical if we want to do great analysis. It is also critical if we want to make sure we’re answering the questions that will lead us to understanding and help us drive meaningful action.

      Turning data into information

      While data is often a starting point in an analytical workflow, it has been said many times that data does not equal information. You can be data rich and information poor. You can have a lot of data and know nothing at all. In fact, you might even say that sometimes the more data you have, the less you know.

      Imagine you have been handed a spreadsheet with a hundred thousand records on it and asked to synthesize it into some useful information.

      Long spreadsheet with many numbers.

      Making sense of this data by just looking at all the numbers would be a virtually impossible task. The spreadsheet helps you organize the data but does little to help turn the data into information.

      Now imagine that the spreadsheet has a latitude and longitude associated with each of the hundred thousand records. How would you make sense of that spatial data? You could start by plotting those locations on a map…but just as with a spreadsheet, simply displaying your data on a map does not turn it into information.

      Map with many points.

      We do get some information from putting data on a map. We begin to visualize the distribution, which gives us a little information, but in some ways, the more data you have, the less information you will gain from adding it to a map. We might be able to make sense of 10 points, but not 100,000.

      Whenever we look at a

      Enjoying the preview?
      Page 1 of 1