A Holistic Approach to Smart Farming: Predictive Analytics and Automation for Resource-Efficient Agriculture

Agriculture is vital for food security, and technology can greatly improve farming efficiency. This project introduces a Smart Farming System integrating Machine Learningbased crop recommendation, fertilizer recommendation, and crop yield prediction with automated timer-based cattle feeding system. The crop recommendation model identifies the best crop using soil macronutrients (N, P, K), rainfall, temperature, and humidity, while the fertilizer recommendation system suggests optimal fertilizers to enhance yield. The crop yield prediction model estimates production based on environmental parameters, and all insights are delivered through a Flask web application where farmers input their soil and weather data. The cattle feeding system, integrated into the same web platform, enables farmers to schedule feeding times for food and water. When the set time is reached, commands are sent via Bluetooth to an ESP32 microcontroller, which activates the food dispenser or water pump. A moisture sensor checks the water level, ensuring the pump runs only when needed to prevent wastage. Each feeding event logs data and sends an SMS notification to the farmer, confirming that livestock has been fed successfully.