airaware — ML pipeline forecasting campus air quality @ Lewis University. airaware-lewis.vercel.app.
AirAware is a real-time machine learning pipeline that forecasts campus air quality at Lewis University — predicting pollution levels before they occur and surfacing actionable insights for campus health and operations decisions.
context:
Lewis University sits in Romeoville, IL, in the Chicago metropolitan area — one of the most air-quality-variable regions in the midwest, affected by industrial output, traffic corridors, and seasonal atmospheric inversions. traditional air quality monitoring is reactive: it tells you what the air is now, not what it's going to be in an hour. AirAware changes that.
the pipeline:
sensor data is ingested from campus-deployed air quality monitors measuring PM2.5, NO2, O3, and CO2. features are engineered from weather data — wind speed, temperature, humidity, precipitation — and temporal features like time-of-day and day-of-week. an ML model (XGBoost ensemble and LSTM for time-series forecasting) is trained on historical sensor readings. forecasts update the dashboard in real time.
what it enables:
predictive alerts for campus administration when air quality is projected to exceed health thresholds. a data foundation for research publications. a framework extensible to other campuses or municipal monitoring networks. the long-term goal is a system that makes environmental health data as accessible and actionable as a weather forecast.