Open-source wildfire prediction & early detection — free for firefighters and everyone at risk.
⚠️ Project status: early scaffold. This repository lays out the structure, mission, and interfaces for Fire Spotter. The risk-modeling code is stubbed and documented but not yet implemented — contributions welcome.
Wildfires move faster than the systems meant to warn us, and the earliest minutes decide everything. Yet the tools to read fire risk are often locked behind agencies, paywalls, or jargon.
Fire Spotter puts that capability in the open — free for the volunteer fire department, the rural family, and the researcher alike. When lives and homes are on the line, the logic that predicts danger should be something anyone can inspect, trust, and improve.
- Wildfire risk scoring — a clear, location-based risk level built from weather, fuel, and terrain inputs.
- Risk mapping — visualize danger across an area to see where conditions are most severe.
- Early-detection signals — surface the earliest indicators of an ignition so crews can respond fast.
- Threshold alerts — notify when conditions cross into dangerous territory.
- Open data inputs — built to work with public weather and environmental data sources.
- Self-hostable — run it on your own infrastructure, no dependence on a paid service.
weather + wind + humidity + drought ┌────────────┐
vegetation / fuel load ─────────► │ Risk model │ ─────► risk score + map + alerts
terrain / slope / aspect └────────────┘
- Read conditions — ingest weather, fuel, and terrain data for an area.
- Model the risk — combine signals into a wildfire-risk score.
- Map & alert — render risk on a map and flag hotspots.
- Spot early — surface early-warning signals while an ignition is still small.
See docs/ARCHITECTURE.md for the full design.
git clone https://github.com/BrettGordon21/firespotter.git
cd firespotter
python3 -m venv .venv && source .venv/bin/activate
pip install -r requirements.txt
python examples/quickstart.pyRight now examples/quickstart.py calls the documented (stub) API and prints a placeholder result, so you can see the intended shape of the project.
- Define the risk-score formula and document its inputs
- Connect a public weather data source (e.g. NOAA / open APIs)
- Add fuel/vegetation and terrain data layers
- Implement
score_risk()for real - Risk map output (GeoJSON / simple web view)
- Threshold alerts
- Validation against historical fire events
Fire Spotter is open source because safety shouldn't be proprietary. Bug reports, data sources, model improvements, and validation help are all welcome — see CONTRIBUTING.md.
MIT — free to use, modify, self-host, and build on. Especially for the volunteer and underfunded departments who need it most.
Built by Brett Gordon · Infinite Space Technologies · Key West, FL