NFL Eras & Team Performance
Spring 2025
NFL Total Score Matrix
Analysis & Results
PCA-Based Season Clustering
Used principal component analysis to reduce season-level scoring data to 9 components, then clustered NFL seasons based on similarities in team scoring patterns. The resulting 5 clusters did not reveal clear, chronologically distinct eras.
Average points scored between team matchups for a given season
For my graduate Matrix Methods course, I worked with a partner to analyze NFL game and performance data using principal component analysis, clustering, and least-squares modeling. We investigated whether distinct eras of NFL play could be identified from historical data and developed models to evaluate offensive, defensive, and overall team performance across seasons.
While the analysis did not reveal clear league-wide eras, our models produced within-season team rankings and highlighted longer-term performance trends across teams.
— Data Analysis · Matrix Methods · Modeling —
Least-Squares Modeling of League Trends
Developed a least-squares model relating team scoring to rushing, passing, and defensive performance, then tracked the fitted coefficients over time to investigate changes in league-wide play.
Teams Performance Over Time
Visualized the passing coefficient from the least-squares model across teams and seasons, revealing periods of sustained passing performance that align with several prominent quarterback tenures.