🌱 AgroPredict AI Platform

AgroPredict AI is a full-stack, AI-powered smart agriculture platform designed to help farmers and agricultural planners make data-driven decisions. By analyzing real-time geolocation-based environmental data and tapping into historical climate patterns, our machine learning models accurately predict crop yields and recommend the most optimal crops to plant.

🌐 Live Demo: https://agro-ml-liart.vercel.app

✨ Key Features

• 🌍 Real-Time Location & Weather Data: Automatically fetches live coordinates using your device's GPS and retrieves local environmental data (rainfall, humidity, temperature) via NASA POWER Climatology and Open-Meteo APIs.
• 📈 Advanced Yield Prediction: Utilizes a highly trained XGBoost Regressor to predict estimated crop yields (Tons/Hectare) based on soil conditions, location data, and weather.
• 💡 Intelligent Crop Recommendations: A multi-class Random Forest / XGBoost Classifier suggests the top crops most suited for your exact geospatial climate and season.
• 🔍 Explainable AI (XAI): Gives transparent insights showing exactly which environmental features (like Rainfall or Soil Type) influenced the AI's decision the most using feature importance graphs.
• 🎨 Modern, Premium UI: A beautiful, responsive frontend built with React, Vite, Tailwind CSS, and Framer Motion, featuring rich shadows, dynamic transitions, and polished typography (Inter & Poppins).

---

📂 Project Structure

This project is separated into a frontend client and a backend API:

• /frontend: The React + Vite SPA. Handles all UI, routing, and animations.
• /backend: The Python FastAPI backend. Houses the Machine Learning models, encoders, datasets, and REST endpoints.

---

🚀 Quick Start Guide

To run the full application locally, you will need to start two separate terminal instances: one for the backend server and one for the frontend development server.

1. Start the Backend API (Terminal 1)

cd backend
python -m venv venv
# Activate the virtual environment
# Windows: venv\Scripts\activate
# Mac/Linux: source venv/bin/activate

pip install -r requirements.txt
uvicorn main:app --reload

The backend API will run on http://localhost:8000.

2. Start the Frontend Client (Terminal 2)

cd "frontend/AgroPredict AI Platform"
npm install
npm run dev

The frontend will run on http://localhost:5173. Open this URL in your browser to interact with the platform.

---

🛠️ Technology Stack

Frontend

• Framework: React.js + TypeScript (managed by Vite)
• Styling: Tailwind CSS + custom CSS variables
• Animations: framer-motion
• Icons: lucide-react
• Charts: recharts

Backend & Machine Learning

• Framework: FastAPI (Python) + Uvicorn
• Machine Learning: XGBoost, Scikit-learn, Pandas, NumPy
• External Data Providers: Open-Meteo API, NASA POWER