AirAware @ Lewis University
at Lewis University, I conduct data-driven evaluations for sensor placement and develop ML-based forecasting models to enhance air quality monitoring effectiveness across campus — translating environmental data into actionable predictions.
research context
Lewis University sits in Romeoville, IL, within the Chicago metropolitan corridor — a region with significant air quality variability from industrial output, highway traffic, and seasonal atmospheric inversions. the research question: can we predict air quality degradation events before they occur, rather than just measuring them after the fact?
the pipeline
sensor data from campus air quality monitors (PM2.5, NO2, O3, CO2) is ingested and combined with weather features — wind speed, temperature, humidity, precipitation — and temporal features. forecasting models including XGBoost ensemble and LSTM networks are trained on historical readings and evaluated against held-out data. the pipeline pushes predictions to a real-time dashboard.
research contributions
contributing to sensor placement optimization using data-driven evaluation of coverage and redundancy. applying coding and statistical techniques to improve data workflow efficiency. participating in research presentations and supporting lab publications through accurate data management and methodology documentation.