EcoNaut
Forecasting harmful algal blooms in the Pacific, in near real time.
- Problem
- My inspiration began while surfing - the water didn't smell right. I later found out it was due to high levels of bacteria from an algal bloom. These blooms can harm marine ecosystems, so I wanted to build something that could help people and wildlife avoid them.
- Approach
- Ingested NASA MODIS ocean-color imagery and NOAA buoy telemetry into PostgreSQL, engineered lag and rolling-window features, trained an XGBoost classifier for bloom events and a Prophet model for SST drift, and served it behind a Flask API feeding a React dashboard.
Data fused
Satellite + Ocean Buoys
Forecast horizon
7 days
Pacific regions
Live dashboard
Solo build · On-going
TypeScriptReactPythonFlaskPostgreSQLXGBoostProphetTime-Series Forecasting
