NBA Player Performance Prediction
A machine learning project to predict NBA players’ next-season performance using historical statistics, trends, and advanced analytics.
Problem
NBA teams invest millions in player contracts, but predicting how a player will perform next season is difficult due to injuries, age, and performance volatility. This project aims to use historical player data to predict future performance and identify players likely to improve or decline.
Dataset
The dataset includes multiple seasons of NBA player statistics such as points, assists, rebounds, minutes played, shooting efficiency, and advanced metrics. Each row represents a player’s performance in a given season.
Feature Engineering
- Rolling averages of points, assists, and rebounds
- Player age and experience
- Minutes per game and usage rate
- Year-over-year performance trends
Model
Several regression models were tested, including linear regression, random forest, and gradient boosting. The final model was selected based on predictive accuracy and stability across seasons.
Evaluation
The model was evaluated using cross-validation and metrics such as RMSE and R² to measure how well it predicted future season performance.
Results
The model successfully identified players likely to improve in scoring and efficiency, as well as players whose performance was expected to decline, helping teams make more informed roster decisions.
What I Learned
This project demonstrated how machine learning can be used to support sports analytics and decision-making. It also highlighted the importance of feature engineering and historical trends when forecasting player performance.