Inspiration:- Farmers often rely on guesswork to estimate yield, fertilizer usage, and irrigation timing. With unpredictable weather patterns, this leads to reduced productivity and resource waste. We built AgriAi4Farmer to bring data-driven precision agriculture to farmers in a simple, accessible way.
What it does:- AgriAi4Farmer is a web-based platform where farmers enter their location, crop, plot size, sowing date, soil type, and water source.
The system instantly provides:
Expected crop yield 2.Recommended NPK fertilizer mix
3.A 7-day irrigation plan
How we built it:- We developed a hybrid ML stack combining: XGBoost for district-level yield patterns ARIMAX-LSTM for dynamic weather and NDVI-based predictions Ridge Regression to reduce bias FAO-56 evapotranspiration model (pyfao56) + Random Forest residual correction for irrigation scheduling We integrated 30+ years of historical weather data, crop yield records, NASA-POWER datasets, and satellite NDVI layers.
Challenges we ran into:- 1.Synchronizing seasonal data to prevent look-ahead bias 2.Integrating multi-source datasets 3.Balancing model complexity with user simplicity
Accomplishments that we're proud of:- 1.Built a robust hybrid AI prediction system 2.Delivered field-specific yield estimates 3.Created a practical, farmer-friendly prototype
What we learned:- Accurate agricultural AI requires strong data alignment, bias prevention, and simplicity in delivery.
What's next for AgriAi4Farmer:- 1.Real-time weather integration 2.Village-level predictions 3.Mobile app deployment 4.Disease prediction modules
Log in or sign up for Devpost to join the conversation.