Dataset link: https://data.sanjoseca.gov/dataset/police-calls-for-service/resource/df207219-ba82-407d-8190-5b31edaded79
- Built a machine learning pipeline using Python, applying K-Means clustering and Random Forest classification to analyze 770,000+ 911 calls and predict crime priority levels with 78% accuracy
- Utilized Pandas, NumPy, and Scikit-learn for feature engineering, data preprocessing, and model evaluation, driving actionable insights from large-scale crime datasets.
- Developed geospatial heatmaps and cluster visualizations, identifying crime hotspots and temporal trends to optimize resource allocation for law enforcement.