FSU Crime Log Map

An interactive, daily-updated map of incidents from the FSU Police Department crime log.

Live site: https://upishanker.github.io/fsu-crime/

• 🗺️ Clustered incident map (Leaflet) with garnet/gold FSU theming and dark/light modes
• 🔎 Filter by category, disposition, and date range, plus free-text search
• 📋 Sortable, paginated incident table
• 📊 Stats dashboard — incidents by category, disposition breakdown, monthly trend
• 🤖 Refreshes automatically every day via GitHub Actions (no server, no API keys)

How it works

The FSU crime log only exposes a rolling ~60-day window, so a daily pipeline accumulates history instead of overwriting it.

scrape → merge into master (dedupe by charge identity) → geocode new addresses (cached) → docs/data/crimes.json → static site

Piece	Location
Data pipeline	pipeline/ (scrape → store → geocode → build)
Master dataset (full history)	data/crime_data.csv
Geocode cache (Nominatim/OSM)	data/geocode_cache.json
Site data artifact	docs/data/crimes.json
Static front-end	docs/{index.html,styles.css,app.js}
Daily automation	.github/workflows/update.yml

Run the pipeline locally

python -m venv .venv && source .venv/bin/activate
pip install -r requirements.txt
python -m pipeline.build          # scrape, merge, geocode, write docs/data/crimes.json
python -m http.server -d docs     # then open http://localhost:8000

Re-running pipeline.build is idempotent: with no new incidents it makes no changes and no geocoding requests.

Geocoding uses the free OpenStreetMap Nominatim service (max 1 req/sec, only for addresses not already cached). The original Google-geocoded data was migrated into the cache via python -m scripts.migrate_seed.