HelloFarmer: An Intelligent Web-Based Agricultural Decision Support Platform for Climate Prediction, Crop Insights, and Market Price Forecasting

Abstract

Agriculture remains one of the most critical yet technologically underserved sectors globally, accounting for the livelihoods of over 1.3 billion people. Farmers face mounting challenges from unpredictable climate patterns, volatile commodity markets, lack of timely crop-specific guidance, and limited access to data-driven tools. Traditional approaches to obtaining weather forecasts, crop advisories, and price information are fragmented, time-consuming, and often inaccessible to smallholder farmers in rural regions. This paper presents HelloFarmer, a comprehensive web-based agricultural decision support platform that integrates three core intelligence modules into a unified, accessible interface: (1) a climate prediction engine that forecasts temperature, rainfall, and humidity using LSTM-based deep learning and Random Forest models trained on historical meteorological data; (2) a crop insight engine that generates personalized crop recommendations and growing advisories by combining real-time climate forecasts with soil parameter inputs using weighted multi-feature scoring; and (3) a market price forecasting module that leverages XGBoost and Facebook Prophet models trained on historical mandi price records to predict near-term commodity prices with quantified confidence intervals. HelloFarmer is built on a Vue.js 3 frontend, a FastAPI asynchronous backend, and a PostgreSQL database, ensuring responsiveness and scalability. Experimental evaluation on Maharashtra regional datasets demonstrates climate forecast MAE below 2°C, crop recommendation top-3 accuracy of 89.2%, and price prediction R² scores exceeding 0.86 across multiple commodity-market pairs. User evaluation with 30 farmers and agricultural students confirmed high task completion rates without prior training, validating the platform’s core objective of democratizing agricultural intelligence for non-expert users.