PurpleAir Temperature Sensor Calibration
Abstract
A machine learning framework for calibrating PurpleAir temperature sensors using a single national XGBoost model with temporal features. The final 63-feature model achieves mean absolute error of 0.64°C and R-squared of 0.992 on the held-out test set. The study includes 797,744 hourly observations from 31 U.S. states.
Fields of Study & Subjects
Air Quality Monitoring and Forecasting
Urban Heat Island Mitigation
Advanced Sensor and Control Systems
Publication details
- Journal / source
- Zenodo (CERN European Organization for Nuclear Research)
- Publication date
- 3 February 2026