Project: Yelp Restaurant Score Predictor & Spatial Clustering

• Applied k-means unsupervised learning to cluster restaurant locations and reveal neighborhood patterns in Berkeley
• Built linear regression-based personalized recommender; computed best single-feature predictor per user via R² maximization, enabling score inference for unrated restaurants
• Generated interactive Voronoi-based visualizations shaded by predicted scores (1–5 stars), highlighting user taste preferences