SmartAgroAI: Multi-Agent System for Sustainable Agriculture

Accenture Hackathon Project – 2025

SmartAgroAI is a full-stack AI-powered multi-agent system designed to optimize sustainable farming practices.

              It brings together farmers, market researchers, and weather stations using intelligent recommendations that reduce 
              environmental impact—like water consumption, carbon footprint, and soil erosion.

Project Structure

. ├── backend/ │ ├── data/ │ │ ├── farmer_advisor_dataset.xlsx │ │ └── market_researcher_dataset.xlsx │ ├── server.py # Main Flask backend logic │ ├── model.py # Crop prediction logic and model loading │ ├── database.py # Interfacing with datasets │ └── weather_api.py # External API integration for weather │ ├── frontend/ │ ├── css/ │ │ ├── styles.css │ │ ├── market.css │ │ └── recommendation.css │ ├── js/ │ │ ├── api-service.js │ │ ├── dataService.js │ │ ├── recommendations.js │ │ ├── config.js │ │ └── connection-check.js │ ├── index.html │ ├── market.html │ └── recommendations.html │ ├── model.pkl # Trained ML model (Random Forest) ├── soil_encoder.pkl # LabelEncoder for soil_type ├── weather_encoder.pkl # LabelEncoder for weather ├── requirements.txt └── README.md """)

How It Works

• ✅ Trained a Random Forest Classifier on farmer and market datasets (.xlsx) using soil, water, and weather data.
• ✅ Loaded encoders and model in model.py for accurate crop predictions.
• ✅ server.py uses Flask to serve the frontend and provide real-time crop & soil treatment recommendations.
• ✅ Weather data is pulled from OpenWeatherMap API (weather_api.py).
• ✅ The frontend connects via api-service.js and connection-check.js to fetch results dynamically. """)

Demo Preview

doc.add_heading('🚀 Demo Preview', level=2) doc.add_paragraph(""" Backend Running Log: INFO – Creating sample farm data INFO – Creating sample market data INFO – Structure verification passed! INFO – Starting Flask server… INFO – Dataset path: d:\Accenture Hackathon\backend\data INFO – Weather API Key Loaded ✅ Flask app running on http://localhost:5000 """)

How to Run Locally

doc.add_heading('🔧 How to Run Locally', level=2) doc.add_paragraph("""

1. Clone the repository git clone https://github.com/your-username/smartagroai.git cd smartagroai

2. Set up the virtual environment python -m venv venv source venv/bin/activate # On Windows: venv\Scripts\activate

3. Install dependencies pip install -r requirements.txt

4. Run the Flask backend cd backend python server.py

5. Access Frontend Open index.html from the frontend/ directory in your browser. """)

API Routes

doc.add_heading('🌐 API Routes', level=2) doc.add_paragraph("""

• /predict_crop – Predicts crop based on soil, water, and weather input
• /get_weather – Fetches current weather
• /get_crop_analysis – Returns performance insights """)

Dependencies

doc.add_heading('📌 Dependencies', level=2) doc.add_paragraph(""" Flask==2.0.1 scikit-learn==1.1.3 pandas==1.5.3 joblib==1.2.0 openpyxl==3.1.2 requests==2.31.0 flask-cors==3.0.10 """)

AI Focus Areas

doc.add_heading('🧠 AI Focus Areas', level=2) doc.add_paragraph("""

• Crop prediction based on soil moisture and weather
• Soil treatment and irrigation schedule suggestions
• Yield increase & water-saving insights
• Future scalability for pest control and CO₂ monitoring """)

Preview

doc.add_heading('📸 Preview', level=2) doc.add_paragraph(""" 📷 Screenshot: server.py log running with model and dataset loaded from /backend/data folder. ✔️ All files connected, Flask API working, weather API successfully called. """)

Status

doc.add_heading('🏁 Status', level=2) doc.add_paragraph(""" ✔️ server.py successfully connects with frontend ✔️ Weather API functional ✔️ Crop recommendations rendering in UI ✔️ Model + encoders (.pkl) are working with live inputs """)

Team

doc.add_heading('🤝 Team', level=2) doc.add_paragraph("""

• Stephen Rodrick – ML Engineer & Full Stack Developer
• Swati Gupta- Frontend Developer

License

doc.add_heading('📩 License', level=2) doc.add_paragraph("This project is under the MIT License.")