Soil Sense - A Nutrient Recommendation System

Abstract

The agricultural sector in the Tamilnadu region, particularly for crops like Tapioca and Turmeric, lacks accessible tools for interpreting technical soil nutrient reports. This paper presents SoilSense AI, a bilingual web-based diagnostic system designed to automate soil health analysis. By integrating real-time data from the Tamil Mannvalam government portal via land survey numbers, the system eliminates manual entry errors. The application employs a Random Forest Regressor algorithm to perform gap analysis and yield prediction based on Nitrogen (N), Phosphorus (P), and Potassium (K) levels. The diagnostic engine translates complex chemical deficits into actionable, split-dosage fertilizer recommendations in both Tamil and English. Experimental results indicate that localized precision nutrient management significantly reduces input costs while enhancing long-term soil productivity. This tool provides a scalable framework for precision agriculture in rural farming communities.