Football Analytics with Python and R: Learning Data Science Through the Lens of Sports

Preface ix1. Football Analytics....1Baseball Has the Three True Outcomes: Does Football?....3Do Running Backs Matter?....4How Data Can Help Us Contextualize Passing Statistics....5Can You Beat the Odds?....5Do Teams Beat the Draft?....6Tools for Football Analytics....6First Steps in Python and R....8Example Data: Who Throws Deep?....10nflfastR in R....11nfl_data_py in Python....14Data Science Tools Used in This Chapter....16Suggested Readings....172. Exploratory Data Analysis: Stable Versus Unstable Quarterback Statistics....19Defining Questions....21Obtaining and Filtering Data....22Summarizing Data....25Plotting Data....29Histograms....30Boxplots....35Player-Level Stability of Passing Yards per Attempt....37Deep Passes Versus Short Passes....41So, What Should We Do with This Insight?....51Data Science Tools Used in This Chapter....52Exercises....53Suggested Readings....533. Simple Linear Regression: Rushing Yards Over Expected....55Exploratory Data Analysis....58Simple Linear Regression....64Who Was the Best in RYOE?....69Is RYOE a Better Metric?....73Data Science Tools Used in This Chapter....76Exercises....76Suggested Readings....774. Multiple Regression: Rushing Yards Over Expected....79Definition of Multiple Linear Regression....79Exploratory Data Analysis....82Applying Multiple Linear Regression....94Analyzing RYOE....100So, Do Running Backs Matter?....105Assumption of Linearity....108Data Science Tools Used in This Chapter....111Exercises....111Suggested Readings....1125. Generalized Linear Models: Completion Percentage over Expected....113Generalized Linear Models....117Building a GLM....118GLM Application to Completion Percentage....121Is CPOE More Stable Than Completion Percentage?....128A Question About Residual Metrics....131A Brief Primer on Odds Ratios....132Data Science Tools Used in This Chapter....134Exercises....134Suggested Readings....1346. Using Data Science for Sports Betting: Poisson Regression and Passing Touchdowns....137The Main Markets in Football....138Application of Poisson Regression: Prop Markets....140The Poisson Distribution....141Individual Player Markets and Modeling....149Poisson Regression Coefficients....162Closing Thoughts on GLMs....169Data Science Tools Used in This Chapter....170Exercises....170Suggested Readings....1717. Web Scraping: Obtaining and Analyzing Draft Picks....173Web Scraping with Python....174Web Scraping in R....179Analyzing the NFL Draft....182The Jets/Colts 2018 Trade Evaluated....192Are Some Teams Better at Drafting Players Than Others?....194Data Science Tools Used in This Chapter....201Exercises....201Suggested Readings....2028. Principal Component Analysis and Clustering: Player Attributes....203Web Scraping and Visualizing NFL Scouting Combine Data....205Introduction to PCA....217PCA on All Data....221Clustering Combine Data....230Clustering Combine Data in Python....230Clustering Combine Data in R....233Closing Thoughts on Clustering....236Data Science Tools Used in This Chapter....237Exercises....237Suggested Readings....2389. Advanced Tools and Next Steps....239Advanced Modeling Tools....240Time Series Analysis....241Multivariate Statistics Beyond PCA....241Quantile Regression....242Bayesian Statistics and Hierarchical Models....242Survival Analysis/Time-to-Event....245Bayesian Networks/Structural Equation Modeling....246Machine Learning....246Command Line Tools....246Bash Example....248Suggested Readings for bash....250Version Control....250Git....251GitHub and GitLab....252GitHub Web Pages and Résumés....253Suggested Reading for Git....253Style Guides and Linting....254Packages....255Suggested Readings for Packages....255Computer Environments....255Interactives and Report Tools to Share Data....256Artificial Intelligence Tools....257Conclusion....258A. Python and R Basics....261B. Summary Statistics and Data Wrangling: Passing the Ball....269C. Data-Wrangling Fundamentals....287Glossary....309Index....317