*Special thanks to Kuan-cheng Fu for providing huge help on this project

In this research I worked with a subset of pitch-level data. The goal was to develop a framework capable of predicting pitch types. The pitch types include Fastball , Curveball, Slider and Changeup. I basically developed a framework including preprocessing features, building predictive models, tuning hyperparameters, evaluating the models, and calculating feature importance. Although the result might not be perfect, I believed that I have provided a reliable framework to deal with this kind of issue. Also it further proved that pitch speed, movement and spin rate are features that we can use to distinguish one pitch kind from another.