A university student in Bangladesh has developed an AI-based crop recommendation system that analyzes soil and weather data to suggest which crops farmers should plant. The system, built by Nahid Islam of Hajee Mohammad Danesh Science and Technology University (HSTU) in Dinajpur, was tested at a local Bangladesh Agricultural Development Corporation (BADC) field with more than 97 percent accuracy, according to a university press release.
The technology combines an IoT soil analysis device, a machine-learning recommendation engine, a Bangla-language chatbot, and an Android app. Islam designed it specifically to be low-cost and accessible to farmers in remote areas, reducing their reliance on sending soil samples to laboratories.
A four-layer system built from local data
Islam, a student in HSTU's electronics and communication engineering department, collected soil samples from the university campus and documented soil characteristics for 28 crop varieties. He then built a machine-learning pipeline covering data preprocessing, feature engineering, model training, and validation.
According to the university, the system works in four layers:
- An IoT-based soil analysis device that measures pH, nitrogen, phosphorus, potassium, moisture, and temperature, transmitting data directly to the cloud.
- A crop recommendation engine that matches field conditions against its database of 28 crops.
- A chatbot built by Islam that answers farmers' questions in Bangla without relying on expensive third-party AI APIs.
- An Android application that lets farmers access recommendations through an ordinary smartphone.
Field testing and academic backing
The researcher recently tested the technology at BADC's nursery field in Dinajpur, where the AI model exceeded the 97 percent accuracy threshold. Agriculturist Shahana Parvin of BADC assisted with the field-level implementation.
The research was supervised by Professor Dr. Dulal Hasan of HSTU's electronics and communication engineering department. "As a teacher, I am truly proud to see this. The system is a practical solution for farmers in the context of Bangladesh," he said in the release.
Why this matters for PR and communications professionals
Technology journalism in Bangladesh and South Asia often focuses on urban startups and export-oriented software. This project shows that students building low-cost, hyperlocal AI tools for agriculture can generate credible stories about national food security and rural development.
When preparing materials for tech clients, note what made this story strong: independent data collection, a clear field trial with measurable results, and a working device farmers could see in person. The chatbot's independence from paid AI services and the availability of the app on a standard Android device, rather than specialized hardware, provided concrete, verifiable details. Those specific elements, not broad claims about AI's benefits, are the ones reporters can piece together into a narrative a reader can trust.
Your membership also unlocks:
