Included in this repo is a project that uses the R package 'nflfastr' to obtain play-by-play data of NFL games. The data was then used to create two models - logistic regression and random forest - that calculate win probabilities throughout the game for each team. Accompanying graphics were also created to visualize win probabilities at different points throughout the game.
Be sure to check out the write-up in 'documentation' for a detailed presentation of methods, results, and analysis!
Please note that due to the size restrictions on Github, the play-by-play file used in the analysis is NOT included in this repo. However, the analysis can be replicated if the user extracts the play-by-play data and game result data from 'nflfastr' directly on their personal computer.