SMARTPADDY: UTILIZING IOT AND ML FOR YIELD FACTOR ANALYSIS

Abstract

Agriculture remains a cornerstone of economic development in many countries, with paddy being a staple crop requiring careful soil nutrient management. Traditional soil testing methods are often slow, expensive, and inaccessible to small-scale farmers, leading to inefficient fertilizer use and declining soil fertility [2][10]. This project introduces a smart, real-time, and cost- effective system that utilizes IoT and machine learning (ML) to monitor and improve the productivity of paddy cultivation [1][3].

The proposed system combines an ESP8266 NodeMCU microcontroller and a 7-in-1 NPK soil sensor to gather crucial soil parameters including nitrogen (N), phosphorus (P), potassium (K), pH, moisture, temperature, and electrical conductivity (EC) [5][7]. These readings are transmitted to the ThingSpeak IoT cloud for real-time monitoring and analysis [2][8]. A machine learning model—developed using the Random Forest algorithm—is hosted on a Flask-based web application to predict soil fertility levels as High, Medium, or Low [3][4][6]. Based on the output, tailored soil reclamation recommendations are provided to enhance fertility [7][9].

The system demonstrated a prediction accuracy of 87% using data collected from Krishi Vigyan Kendra (KVK) and field experiments [10]. It offers actionable, crop-specific guidance that enables timely intervention, thereby reducing wasteful fertilizer use and increasing yield [1][4]. With its user-friendly interface and low-cost design, this solution is particularly beneficial for farmers in rural and resource-limited settings [2][5].

This research highlights the potential of integrating embedded systems, cloud IoT platforms, and ML to build intelligent, scalable agricultural tools that contribute to sustainable farming and improved food security [6][8].