About

Machine learning project to provide predictions for all the seasonal NBA Awards.

Technical Details

• All data (respectfully) scraped from BasketballReference.com.
• All models trained on aggregated box score data (games started, minutes, percentage of team games played, points, shooting %s, rebounds, assists, steals, blocks, plus minus etc.).
• Models trained using CatBoost.
• Models / Predictions / SHAP explanations saved to PostgreSQL.
• Django / Boostrap used for web interface.
• CI/CD via Drone & gitea
• Dokku for serving on a Hetzner VPS.

Contact: @dougdeep7 with any questions or comments!