Welcome
The present project is an air quality monitoring station that has been assembled using a set of MQ gas sensors, a module to monitor temperature and humidity, and a module for the detection of sound. All the data collected by this air quality monitoring station is stored in a database.
Use the Data page to explore live and historical sensor readings with interactive charts, filtered by date and location. The Predictions page surfaces machine-learning forecasts derived from the collected data. Registered users can also download the full dataset as a CSV.
About the station
AirNest is built around a Raspberry Pi 4B and monitors indoor air quality using five MQ-series gas sensors: the MQ-135 (equivalent CO₂ / general air quality), MQ-3 (alcohol and benzene), MQ-6 (LPG and butane), MQ-7 (carbon monoxide), and MQ-8 (hydrogen). A DHT11 module records temperature and humidity, and a KY-037 sensor tracks ambient sound levels and sudden sound events.
Readings are taken every 30 minutes and sent to a cloud database. The Predictions page shows a daily 7-day forecast generated by two machine learning models: Prophet, used for temperature and humidity forecasting, and a Random Forest classifier, which predicts whether air quality conditions are likely to reach a danger level based on historical sensor data and weather forecasts from Open-Meteo.