Urban Safety Analytics: San Francisco Crime Trend Analysis (2018-2025)
This project examines San Francisco crime trends from 2018 to 2025 using time-series analysis and spatial visualization techniques. Incident-level data were aggregated by time, category, and neighborhood to identify seasonal patterns, trend shifts, and emerging hotspots. Forecasting models were applied to project future crime levels under observed conditions. Results highlight disparities across neighborhoods and provide evidence-based insights relevant to public safety strategy, resource allocation, and policy evaluation.
Problem
San Francisco crime data is large, time-dependent, and difficult to translate into actionable insights for operational decision-makers.
Data
San Francisco Police Department open datasets covering reported incidents from 2018 to 2025, cleaned, standardized, and organized by neighborhood.
Methods
Exploratory data analysis, feature engineering, time-series modeling (ARIMA/SARIMA), spatial pattern analysis, and interactive visualization.
Outcome
Identified high-risk neighborhoods, forecasted crime trends, and delivered interactive dashboards to support patrol planning and resource allocation.
Tools
Python, Pandas, NumPy, Statsmodels, GeoPandas, Tableau, Streamlit, GitHub.