Crop Prediction
An IoT-driven farming portal powered by machine learning algorithms that recommend optimal crop selection based on dynamic NPK soil sensors and atmospheric logs.
95.4%
Scikit-Learn Random Forest model
Realtime
CoAP/MQTT sensor telemetry
3 Channels
Nitrogen, Phosphorus, Potassium logs
<85ms
Hosted Django REST endpoint
IoT Soil Dashboard Mockup
42 mg/kg
28 mg/kg
50 mg/kg
28.4°C
62%
Recommended Crop: Maize
Optimal soil and climatic compatibility detected.
Algorithm & Model Selection
The solution utilizes an ensemble learning algorithm (Random Forest Classifier) trained on historical agricultural datasets compiling soil pH, NPK, temperature, rainfall, and humidity logs. The trained model is served via a Python Django REST API that receives real-time ESP32 micro-controller sensor updates.
Configured ESP32 micro-controller configurations using NPK electrochemical soil probes, compiling sensors logs dynamically.
Designed light-weight Django API endpoints utilizing custom rate-limiting protocols to support high-frequency sensor streams.
Validated recommendation models under cross-validation runs, achieving 95.4% accuracy across 22 distinct crop categories.
Created Flutter cross-platform applications providing visual soil dashboard analytics and graphical trends over historical logs.