My goal was to scrape and analyze data about competitive pokemon then classify pokemon into tiers through machine learning. Recently, I added a notebook analyzing the pokemon tiers themselves, and modelled how a pokemon changes tiers through a markov chain.
The "data collection" folder the scripts used to scrape Smogon, Bulbapedia, and Veekun for the pokemon data. The "K-Modes Models" folder contains pickled versions of the k-modes with a varying number of modes for the move clustering.
You can read about the lessons I learned through this project here: https://nicholasvadivelu.com/2018/09/06/competitive-pokemon/