This project demonstrates one of the many uses of machine learning in baseball by creating a machine learning model capable of identifying which pitch is being thrown when given the pitcher, velocity, and movement of the pitch.
Data used is taken from Baseball Savant.
The notebook file walks through the creation of both a general and pitcher-specific machine learning model, as well as an analysis of the accuracy of both models.
The machine learning model can be interacted with via a Streamlit web app dashboard. A user can select either the general or pitcher-specific model and by entering the pitcher (pitcher-specific model only), pitch velocity and movement, the model will identify and display what pitch was thrown.
Check out the web app here.