NBA Fuzzy Clustering

This repo explores various centroid-based clustering methods (outlined in slides 1-19) and implements a fuzzy clustering algorithm for NBA player classification as well as several other centroid-based clustering methods (slides 20-40).


After using k-means++ to cluster top PGs into two groups using a variety of statistics, we wanted to visualize which individual statistics partioned the PGs best. We see here that 3-point percentage is a clear separator.