AgriClimateBD: A Satellite Enriched Precision Agriculture Dataset for Bangladesh

AgriClimateBD is a comprehensive precision agriculture dataset developed for Bangladesh by integrating agricultural production statistics, crop phenological information, and satellite-derived climate variables. The dataset contains 4,607 records and 33 features representing 73 crop types cultivated across 64 districts of Bangladesh. The dataset was developed by extending the previously published SPAS-Dataset-BD through the incorporation of climate information obtained from the NASA POWER (Prediction of Worldwide Energy Resources) platform. Agricultural production, area, temperature, and humidity information were collected from the Bangladesh Bureau of Statistics (BBS) Statistical Yearbook 2022, while crop lifecycle information, including transplanting, growth, harvesting periods, and seasonal classifications, was obtained through field surveys involving 223 farmers from diverse agroecological regions of Bangladesh. A Python-based automated data enrichment pipeline was employed to retrieve daily climate observations from the NASA POWER API for all 64 districts. Four climate parameters—precipitation, solar radiation, wind speed, and evapotranspiration—were aggregated into phenologically meaningful temporal windows. These include full-season summaries, pre-transplant month, transplant month, and post-transplant month conditions. The resulting climate features were integrated with agricultural and phenological attributes to create a unified dataset suitable for precision agriculture research. The dataset includes agricultural variables such as crop name, district, cultivated area, production, season, transplanting period, growth period, and harvest period; meteorological variables including average, maximum, and minimum temperature and humidity; and sixteen NASA POWER-derived climate variables representing rainfall, solar radiation, wind speed, and evapotranspiration across multiple crop-development stages. AgriClimateBD supports a wide range of applications including crop yield prediction, crop classification, climate-resilient agriculture, agricultural decision support systems, precision irrigation management, machine learning, deep learning, and agricultural policy analysis. The dataset is particularly valuable because it provides phenologically anchored climate information for Bangladesh’s agricultural systems, including numerous underrepresented crop species that are rarely available in public agricultural datasets. The dataset is provided in CSV format and is intended to facilitate reproducible research in agricultural informatics, climate-smart agriculture, remote sensing, and data-driven farming systems.