Built predictive models on NBA games using team stats to predict winners pre- vs post-All-Star break. Compared logistic regression, CART, and random forest models. Used the best model to simulate the 2024 NBA Playoffs.
Built a March Madness prediction model using a Graph Neural Network. Each team was represented as a point and each game as a connection between teams, then trained the model to learn from past matchups and predict future tournament winners.
