Who This Book Is For

Our book has two target audiences. First, we wrote the book for people who want to learn about football analytics by doing football analytics. We share examples and exercises that help you work through the problems you’d face. Throughout these examples and exercises, we show you how we think about football data and then how to analyze the data. You might be a fan who wants to know more about your team, a fantasy football player, somebody who cares about which teams win each week, or an aspiring football data analyst. Second, we wrote this book for people who want an introduction to data science but do not want to learn from classic datasets such as flower measurements from the 1930s or Titanic survivorship tables from 1912. Even if you will be applying data science to widgets at work, at least you can learn using an enjoyable topic like American football.

We assume you have a high school background in math but are maybe a bit rusty (that is to say, you’ve completed a precalculus course). You might be a high school student or somebody who has not had a math course in 30 years. We’ll explain concepts as we go. We also focus on helping you see how football can supply fun math story problems. Our book will help you understand some of the basic skills used daily by football analysts. For fans, this will likely be enough data science skills. For the aspiring football analyst, we hope that our book serves as a springboard for your dreams and lifelong learning.

To help you learn, this book uses public data. This allows you to re-create all our analyses as well as update the datasets for future seasons. For example, we use only data through the 2022 season because this was the last completed season before we finished writing the book. However, the tools we teach you will let you update our examples to include future years. We also show all the data-wrangling methods so that you can see how we format data. Although somewhat tedious at times, learning how to work with data will ultimately give you more freedom: you will not be dependent on others for clean data.

Who This Book Is Not For

We wrote this book for beginners and have included appendixes for people with minimal-to-no prior programming experience. People who have extensive experience with statistics and programming in R or Python would likely not benefit from this book (other than by seeing the kind of introductory problems that exist in football analytics). Instead, they should move on to more advanced books, such as R for Data Science, 2nd edition by Hadley Wickham et al. (O’Reilly, 2023) to learn more about R, or Python for Data Analysis, 3rd edition by Wes McKinney (O’Reilly, 2022) to learn more about Python. Or maybe you want to move into more advanced books on topics we touch upon in this book, such as multivariate statistics, regression analysis, or the Posit Shiny application.

We focus on simple examples rather than complex analysis. Likewise, we focus on simpler, easier-to-understand code rather than the most computationally efficient code. We seek to help you get started quickly and connect with real-world data. To use a quote often attributed to Antoine de Saint-Exupéry:

If you wish to build a ship, do not divide the men into teams and send them to the forest to cut wood. Instead, teach them to long for the vast and endless sea.

Thus, we seek to quickly connect you to football data, hoping this connection will inspire and encourage you to continue learning tools in greater depth.

How We Think About Data and How to Use This Book

We encourage you to work through this book by not only reading the code but also running the code, completing the exercises, and then asking your own football questions with data. Besides working through our examples, feel free to add in your own questions and create your own ideas. Consider creating a blog or GitHub page to showcase your new skills or share what you learn. Work through the book with a friend or two. Help each other when you get stuck and talk about what you find with data. The last step is especially important. We regularly think about and fine-tune how we share our datasets as we work as professional data scientists.

In our day jobs, we help people make decisions using data. In this book, we seek to share our tools and thought processes with you. Our formal academic training covered mathematics and statistics, but we did not truly develop our data science skills until we were required to analyze messy, ecological, and environmental data from the natural world. We had to clean, merge, and wrangle untidy data, all while figuring out what to do with gaps in the information available to us. And then we had to try to explain the meaning hidden within those messy datasets.

During the middle of the last decade, Eric starting applying his skills to football, first as a writer and then as an analyst, for a company called Pro Football Focus (PFF). Eventually, he left academia to join PFF full time, helping run its first data science group. During his time with PFF, he worked with all 32 NFL teams and over 130 college teams, before moving to his new role at SumerSports. Meanwhile, Richard continues to work with ecological data and helps people make decisions with this data—for example, how many fish need to be harvested from where in order to control an undesired species?

Although we both have advanced degrees, the ability to think clearly and explore data is more important than formal education. According to a quote attributed to Albert Einstein, “Imagination is more important than knowledge.” We think this holds true for football analytics as well. Asking the right question and finding a good enough answer is more important than what you currently know. Daily, we see how quantitative tools help us to expand our imagination by increasing our ability to look at and think about different questions and datasets. Thus, we are required to imagine important questions to guide our use of analytics.

A Football Example

Let’s say we want to know if the Green Bay Packers have a home-field advantage. Perhaps we disagree with a friend over whether the Frozen Tundra truly is the advantage that everyone says it is, or if fans are wasting their hard-earned money on an outsized chance of getting frostbite. Conceptually, we take the following steps:

1. We find data to help us answer our question.

2. We wrangle the data into a format to help us answer our question.

3. We explore the data by plotting and calculating summary statistics.

4. We fit a model to help us quantify and confirm our observations.

5. Lastly, and most importantly, we share our results (optionally, but possibly most importantly, we settle the wagers to the “questions” we answered with data).

For the Packers home-field advantage example, here are our concrete steps:

1. We use the nflreadr package to obtain data, which is freely available to use.

2. We wrangle the data to give us a score differential for each game.

3. We create a plot such as that in Figure P-1 to help visualize the data.

4. We use a model to estimate the average difference point differential for home and away games.

5. We share our results with our friends who were debating this topic with us.

Given the data, the Packers typically have a point differential of two points higher at home compared to being on the road. This is in line with where home-field advantage is assumed to be across the league, although this number has evolved substantially over time.

Figure P-1. Green Bay score differential in home and away games from 2016 to 2022

Hopefully, this observation raises more questions for you. For example, how much variability exists around this estimate? What happens if you throw out 2020, when fans were not allowed to attend games at the Packers home stadium? Does home-field advantage affect the first half or second half of games more? How do you adjust for quality of play and schedule differential? How does travel distance affect home-field advantage? Familiarity? Both (since familiarity and proximity are related)? How does weather differential affect home-field advantage? How does home-field advantage affect winning games compared to losing games?

Let’s quickly look at homefield advantage during games won by the Packers compared to games lost. Look at Figure P-2. Summaries for the data in this figure show that the Packers lose by one point more on the road compared to home games, but they win at home by almost six points more than they do on the road.

Figure P-2. Green Bay score differential in home and away games from 2016 to 2022 for winning and losing games

That being said, the Green Bay Packers are a good team and will likely have a positive point differential no matter what. The question of home-field advantage is not a trivial one and has perplexed analysts and bettors alike for decades. Hopefully, this example spurs more questions for you. If so, you’re reading the right book.

We cover step 1, obtaining football data, in Chapter 1. We cover step 2, exploring data, in Chapter 2. We cover step 3, data wrangling, in case studies throughout the book as well as in Appendixes B and C. We cover step 4 with basic statistics in Chapter 2 and Appendix B and then cover models in Chapters 3 through 8. We cover step 5 throughout various chapters as we describe what we have found. Lastly, we round out the book with Chapter 9, which describes some of the advanced tools we use daily.

Conventions Used in This Book

The following typographical conventions are used in this book:

Italic

Signifies new terms, URLs, email addresses, filenames, and file extensions.

Constant width

Used for program listings, as well as within paragraphs to refer to program elements such as variable or function names, databases, data types, environment variables, statements, and keywords.

Constant width bold

Shows commands or other text that should be typed literally by the user.

Constant width italic

Shows text that should be replaced with user-supplied values or by values determined by context.

Using Code Examples

Supplemental material (code examples, exercises, etc.) is available for download at https://github.com/raerickson/football_book_code.

If you have a technical question or a problem using the code examples, please send an email to [email protected].

This book is here to help you get your job done. In general, if example code is offered with this book, you may use it in your programs and documentation. You do not need to contact us for permission unless you’re reproducing a significant portion of the code. For example, writing a program that uses several chunks of code from this book does not require permission. Selling or distributing examples from O’Reilly books does require permission. Answering a question by citing this book and quoting example code does not require permission. Incorporating a significant amount of example code from this book into your product’s documentation does require permission.

We appreciate, but generally do not require, attribution. An attribution usually includes the title, author, publisher, and ISBN. For example: “Football Analytics with Python and R by Eric A. Eager and Richard A. Erickson (O’Reilly). Copyright 2023 Eric A. Eager and Richard A. Erickson, 978-1-492-09962-8.”

If you feel your use of code examples falls outside fair use or the permission given above, feel free to contact us at [email protected].

O’Reilly Online Learning

Our unique network of experts and innovators share their knowledge and expertise through books, articles, and our online learning platform. O’Reilly’s online learning platform gives you on-demand access to live training courses, in-depth learning paths, interactive coding environments, and a vast collection of text and video from O’Reilly and 200+ other publishers. For more information, visit https://oreilly.com.

How to Contact Us

Please address comments and questions concerning this book to the publisher:

• O’Reilly Media, Inc.
• 1005 Gravenstein Highway North
• Sebastopol, CA 95472
• 800-889-8969 (in the United States or Canada)
• 707-829-7019 (international or local)
• 707-829-0104 (fax)
• [email protected]
• https://www.oreilly.com/about/contact.html

We have a web page for this book, where we list errata, examples, and any additional information. You can access this page at https://oreil.ly/football-analytics.

For news and information about our books and courses, visit https://oreilly.com.

Find us on LinkedIn: https://linkedin.com/company/oreilly-media

Follow us on Twitter: https://twitter.com/oreillymedia

Watch us on YouTube: https://youtube.com/oreillymedia

Acknowledgments

We thank our editors at O’Reilly, including Michelle Smith, Corbin Collins, Clare Laylock, and Aleeya Rahman for their support. We also thank Sharon Wilkey for copyediting our book, Larry Baker at Amnet Systems LLC for proofreading, and Cheryl Lenser at nSight, Inc. for creating our index. We thank Nick Adams and Danny Elfanbaum for technical assistance with the O’Reilly Atlas system. We thank Boyan Angelov, Richie Cotton, Matthew Coller, Molly Creagar, Ryan Day, Haley English, Chester Ismay, Kaelen Medeiros, George Mound, John Oliver, and Tobias Zwingmann for technical feedback. We also thank Richie Cotton for tips on writing a successful book proposal.

Eric would like to thank his wife, Stephanie, for her patience and constant support of his dreams, regardless of how crazy. He would also like to thank Neil Hornsby, whose vision in building PFF gave him a platform that has been the foundation for everything he’s done until now, and Thomas Dimitroff, for answering his email back in the fall of 2020. He’d also like to thank Paul and Jack Jones for their vision in starting SumerSports in 2022. Finally, Eric would like to thank his parents, who, despite not being huge football fans themselves, kindled the flames of his passion throughout his childhood.

Figure P-3. Margo (left) and Sadie (right) walking. Who is walking whom? (Photo by Richard Erickson)

Richard thanks his daughter, Margo, for sleeping well so he could write this after she went down for the night. He also thanks Sadie for foregoing walks while he was writing, as well as patiently (and impatiently) reminding him to take stretch breaks (Figure P-3). Richard also thanks his parents for raising him with curiosity and his brother for prodding him to learn to program and for help with this book’s proposal. Lastly, Richard thanks Tom Horvath and others from Hale, Skemp, Hanson, Skemp, & Sleik for support while writing this book.