Appropriate Crop Recommendation Dataset for Cultivation in Bangladesh using IoT and Machine Learning (CropRec-BD v1, 2025)

This dataset supports the study “Appropriate Crop Recommended System for Cultivation using IoT and ML” and was collected in Bangladesh (multiple regions) between 20XX–2025 using low-cost IoT sensors, farmer surveys, and agricultural records. It contains time-series environmental measurements (soil moisture, soil temperature, pH, ambient temperature, relative humidity, light), geo-located field metadata (district, soil type, elevation), crop outcome labels (suitable crop recommendations, historical yields), and contextual socio-economic variables (farm size, irrigation type, planting dates, fertilizer usage). The dataset is organized to facilitate supervised and semi-supervised machine learning for crop recommendation, yield prediction, and interpretability research. Data has been anonymized to remove personally-identifying information; consent was obtained from participating farmers. Example use cases: training crop recommendation classifiers, benchmarking transfer learning across regions, developing explainable AI models for agronomic advisories. The dataset includes raw CSV files, a cleaned/merged master table, feature dictionaries, and sample code (Python notebooks) demonstrating preprocessing and a baseline ML model. Licensed under CC BY 4.0 to enable academic and commercial reuse with attribution

The data was gathered over the past 5 years (2017 to 2022)
of agricultural data from the websites of the Bangladesh
Agricultural Research Institute and NASA. There are
roughly 3,000 records for nine different crops. Temperature,
humidity, and rainfall are all included in the crop-suggested
information.
The databases contain crop data from all of Bangladesh.
Several different types of crops are grown in Bangladesh. It is
challenging to analyse every crop variety. Therefore, we have
done the top 9 crops that are grown extensively. When it comes
to preprocessing, all the data are transformed into numerical data. A similar procedure was used for fertilizer data.
List of crops considered:
• Maize
• Strawberry
• Wheat
• Potato
• Banana
• Mango
• Jute
• Pineapple
• Sugercane