Web Scraping & Linear Regression Modeling for NFL Points Total

Description

Using various traditional and modern metrics to predict the total points scored in an NFL game.

Features & Target Variables

• Target: total points in a single NFL game
• Features:
  • Traditional statistics:
    • Yards
    • Touchdowns
    • First Downs
    • Third Downs
    • Fourth Downs
    • Attempts, Completions, Interceptions
    • Rush attempts, Fumbles
    • Etc
  • Efficiency metrics:
    • Offensive Expected Points Added (EPA)
    • Passing EPA
    • Rushing EPA
    • Turnover EPA
    • Etc

Data Used

• pro-football-reference.com

Tools Used

• Python
• Jupyter notebooks
• Pandas
• Numpy
• Sk-learn
• Beautiful Soup

Possible Applications

This model can be used to evaluate how well a line was set at a sports book, or used to analyze how well a game plan was executed or what critical parts of a team's performance need to be improved or focused on more.