Home
Wildfire Smoke Risk Correlator¶
A local-first research and engineering stack that correlates NASA FIRMS wildfire hotspots and OpenAQ air-quality measurements with U.S. Census geometries using Kafka, Apache Spark, and PostGIS. It publishes smoke-risk indices, optional plume and dispersion-style proxies, calibration scaffolding, and SQL-first alerting.
Non-claims
This software is not a public-health advisory system, not regulatory dispersion modeling, and not operational emergency guidance. Outputs are engineering correlations for experimentation and operator inspection.
Quickstart
Compose, topics, database bootstrap, and first steps.
Architecture
How sources, Kafka, Spark, and PostGIS fit together.
Operations
Bootstrap, Kafka/DLQ, dashboards, calibration exports.
Calibration
Evaluation scaffolding — honest metrics, not validation claims.
Release notes
v1.1.0 documentation site, CI gates, and validation commands.
Who this documentation is for¶
- Researchers and engineers running reproducible demos on open data.
- Operators inspecting Postgres views, Grafana, and alerts.
Start with Quickstart or the No-secrets demo.
Quick links¶
| I want to… | Go to |
|---|---|
| Run fixtures without API keys | No-secrets demo |
| Bring up Docker Compose and bootstrap DB | Quickstart, Database bootstrap |
| Ingest live data (bounded) | Live ingest |
| Inspect outputs / Grafana | Dashboards |
| Understand risk bands and models | Risk models |
| Fix migration drift | Troubleshooting, Database doctor |
Repository: GitHub.