Bayesian Baseball

Masters project presented at the 2019 Joint Statistical Meetings

“Bayesian Baseball” describes and analyzes a multi-level Bayesian regression model which attempts to predict the outcome of a baseball game. Several panels focus on the 2018 World Series between the Boston Red Sox (BOS) and the LA Dodgers (LAD).

The data, which includes game-level (batting, pitching) and team-level (fielding) variables, was scraped from baseball-reference.com

The model was compiled using Stan in R.

The graphs were exported from R Studio as vector images in order to maintain a crisp, scalable image on a large, 4’ x 8’ poster. The poster was assembled using Adobe Photoshop.

The poster was constructed with feedback from my Professor, Ayona Chatterjee, and presented at the 2019 JSM in Denver, CO.