Redefining NBA positions and classifying those for incoming prospects

Project synopsis

The National Basketball Association (NBA) has grown more dynamic and versatile with its maturity. Traditional positions fail to capture roles and play styles, erring more toward legacy physical profiles. Are there natural NBA player groupings that better reflect the current talent pool? Yes! Through clustering and predictive modeling techniques, we've derived 10 new player position classifications for current league players and incoming propsects.

Project Report

Final report detailing the scope of analysis, results from clustering and classification, recommendations, limitations, and future opportunities..

executive summary report pdf

Machine Learning Applications

Processing NBA data for deriving new position clusters, studying their relationships, and classifying positions for incoming prospects.

modeling clustering pca kmeans neural net quarto r
Data Cleaning

Data Wrangling

Cleaning up and joining multiple data extracts from PBPStats dot com.

data wrangling quarto r