AeroSentinel turns fragmented air-quality data into a local, one-hour-ahead PM10 estimate and a separate high-event risk signal.
The platform combines what monitors are measuring now with the conditions that move, trap, wash out or create particulate matter.
A continuous concentration estimate in µg/m³, designed for short-term awareness at monitored locations.
A separate risk score flags when PM10 may cross elevated thresholds—without distorting the everyday forecast.
Forecast ranges and data-quality warnings help users understand when the system is confident and when conditions are changing fast.
The pipeline can be retrained for another county using local monitors, weather, smoke guidance, emissions and a strict held-out test.
Smoke forecasts understand regional transport. Monitor histories understand what is actually happening near the ground. AeroSentinel combines both, then learns the local bias.
We’ll walk through data availability, evaluation and a pilot path.