Kaggle Competition: NFL Big Data Bowl

Project Overview

In this project, I set out to analyze how wide receivers behave while the ball is in the air, a critical but understudied moment in football analytics. Using frame-by-frame tracking data from the NFL Big Data Bowl, I built a new metric called Ball-Tracking Intelligence (BTI) that measures how intelligently and efficiently a WR reacts once the ball is thrown. I engineered five movement-based sub-metrics — alignment, path efficiency, smoothness, early reaction, and late adjustment — and combined them using SHAP-derived weights for interpretability. This allowed me to produce per-play and per-player BTI scores, create leaguewide rankings, adjust for play difficulty, and compare performance across route types. My results show that BTI captures a meaningful, repeatable skill linked to catch success and deep-ball performance, offering valuable insights for scouting, player development, and scheme design. If you want, I can also make a 1-sentence or bullet-point version.

Technologies Used

Impact

Created a metric BTI: Ball Tracking Intelligence to evaluate Wide Receiver's ability to track the ball while in the air