Predict air quality (NO₂)
Guided regression project · 17 steps · Ridge · gradient boosting
About
Hourly multivariate sensor data from the UCI Air Quality repository. Values of -200 mark missing readings. Rows stay in time order: we evaluate with a time-based split so we do not train on the future to predict the past.
Data
CSV at /datasets/air-quality/AirQualityUCI.csv— semicolon-separated, European decimal commas. Target is ground-truth NO2(GT) (µg/m³); features are other pollutant and weather columns (date/time excluded).
Task
Complete TODOs top-to-bottom. Use hints first. Switch to Sample solution only when stuck.
Pro project
Unlock guided projects to solve realistic ML tasks end-to-end, run full notebooks in the browser, and build interview-ready project confidence.
Upgrade to ProDataset credit: Air Quality dataset by S. Vito (UCI Machine Learning Repository, DOI: 10.24432/C59K5F), licensed under CC BY 4.0. Dataset page: UCI Air Quality.