This was my group’s project for the 2023 NFL Big Data Bowl, where we were selected as a finalist and presented our work at the NFL Combine. We utilized a XGBoost model to predict whether a defender would rush the passer based on four features: distance from line of scrimmage, horizontal distance from ball, direction and their distance crept towards the line before the snap. Our model correctly predicts a defender will pass rush 91% of the time while incorrectly predicting a coverage player will pass rush 6% of the time. This project is one of my personal favorites, and our full write up can be found here: xPassRush. Our code for the project can be found here: xPassRush Code.