Delhi,Mumbai,Kolkata,Guwahati PM2.5 Data Set

Processed, model-ready PM₂.₅ datasets for four Indian cities — Delhi, Kolkata, Guwahati, and Mumbai — at a uniform 3-hourly resolution. Each file provides station-level fine particulate matter (PM₂.₅, µg/m³) together with co-located meteorology: air temperature, relative humidity, wind speed, and wind direction (the latter also encoded as sine/cosine components). Delhi covers 40 stations over 2021–2022 (233,600 records) and includes one-hot month and hour-of-day features; Kolkata (10 stations) and Guwahati (4 stations) cover 2023 and additionally include the co-pollutants NH₃, NOₓ, O₃, and PM₁₀ with one-hot season indicators; Mumbai covers 28 stations over 2023–2024 (163,744 records), harmonized to the common 3-hourly grid from hourly observations and serving as a coastal, out-of-airshed control. All measurements were obtained from the Central Pollution Control Board (CPCB), India. The datasets support a cross-city transfer-learning study of PM₂.₅ forecasting across the Indo-Gangetic Plain.