This is the home page for the Pitch Prediction project created as a capstone for the UC Berkeley Masters of Information and Data Science (MIDS) program by Zach Beaver, Jason Goodman, Josh Lu, and Alan Si. The purpose of this project is to buildĀ machine learning models that, given features about the current game state and a pitcher’s past behavior, can predict which pitch will be thrown next.

Information regarding the project can be viewed in the menu bar links, and the output of the model currently has two forms. In the sidebar is a Twitter bot that live-tweets pitch predictions during games, and linked to the image below is an output explorer that allows for various examinations of the model for different inputs and outputs.

Shiny App for Exploring Model Outputs from 50 Randomly Selected Pitchers