CROP MSP Forecasting and OTP-Verified SMS Notification System

Abstract

Agricultural price forecasting plays a vital role in empowering farmers with market intelligence, enhancing crop planning, and supporting economic resilience. This project presents an efficient and user-friendly system for predicting the Minimum Support Price (MSP) of crops using machine learning techniques, with a real-time interface built using Streamlit. The system leverages an XGBoost regression model trained on historical crop price datasets, including commodity name, crop variety, type, and year. To increase accessibility and impact, the application incorporates Twilio SMS integration, enabling users to send MSP predictions directly to farmers’ mobile phones. The web interface includes a step-wise selection mechanism for crop type, commodity, and variety, along with intuitive visualization of prediction results and comparison with actual MSP values when available. The model achieves a strong R² score, indicating reliable predictive performance across crop types and years. By integrating machine learning with SMS-based communication, this solution offers a practical and scalable tool for agricultural advisory systems, especially in rural and low-resource settings.