🌾 Project Samarth

Intelligent Q&A System for Indian Agricultural & Climate Data

🎥 Demo Video

▶️ Watch 2-Minute Demo on Loom

🎯 Overview

AI-powered system that answers natural language questions about Indian agriculture and climate using live data from data.gov.in APIs.

✨ Key Features

• 🔍 Real-time data fetching from government APIs
• 💬 Natural language query processing
• 🤖 AI-powered answers with source citations
• 📊 Automatic query analysis
• 🎨 Clean web interface

📊 Data Sources

• Crop Production API: 246,091 records (1997-2014)
  • District-wise, season-wise crop production statistics
• Daily Rainfall API: 3.4M+ records (2018-present)
  • District-wise daily rainfall data
• Source: data.gov.in

🛠️ Tech Stack

• Python 3.11+
• Streamlit (Web Interface)
• Groq AI (Free LLM)
• Pandas (Data Processing)
• data.gov.in REST APIs

🚀 Quick Start

# Clone repository
git clone https://github.com/priyanka7411/project-samarth.git
cd project-samarth

# Install dependencies
pip install -r requirements.txt

# Set up environment
echo "GROQ_API_KEY=your_key_here" > .env

# Run app
streamlit run app.py

Get free Groq API key: https://console.groq.com/

💡 Example Questions

Try asking:

• "Compare rice production in Punjab and Tamil Nadu for 2013"
• "Show wheat production in Haryana for 2012"
• "What is the rainfall pattern in Kerala for 2020?"
• "Which districts have highest sugarcane production in 2011?"

🏗️ System Architecture

Components:

1. Query Analyzer (query_analyzer.py)

   • Extracts states, crops, years using pattern matching
   • Identifies query type (comparison, trend, ranking)

2. Data Handler (data_handler.py)

   • Manages API pagination and rate limiting
   • Fetches from multiple data.gov.in APIs
   • Handles different data formats

3. AI Engine (ai_system.py)

   • Coordinates data fetching and analysis
   • Uses Groq LLM for natural language generation
   • Ensures all answers cite sources

4. Web Interface (app.py)

   • Streamlit-based UI
   • Query history management
   • Real-time analysis display

┌─────────────────┐
│   User Query    │
└────────┬────────┘
         │
         ▼
┌─────────────────┐
│ Query Analyzer  │  ← Extracts states, crops, years
└────────┬────────┘
         │
         ▼
┌─────────────────┐
│  Data Handler   │  ← Fetches from data.gov.in APIs
└────────┬────────┘
         │
         ▼
┌─────────────────┐
│   AI Engine     │  ← Groq LLM generates answer
└────────┬────────┘
         │
         ▼
┌─────────────────┐
│  Answer + Cites │
└─────────────────┘

Data Flow:

User Query → Query Analysis → Data Fetching → AI Processing → Answer + Citations

🎯 Design Decisions

Real-time API Fetching

• Ensures data accuracy and freshness
• No local storage needed (better for security)
• Can be deployed in air-gapped environments

Free Groq API

• No cost barriers
• Better data sovereignty vs paid services
• Fast inference times

Source Citations

• Every answer includes data.gov.in API references
• Builds trust in AI-generated responses
• Enables verification of claims

Modular Architecture

• Easy to add new data sources
• Separation of concerns
• Testable components

📁 Project Structure

project-samarth/
├── app.py                 # Streamlit web interface
├── ai_system.py          # Main AI Q&A system
├── data_handler.py       # API data fetching logic
├── query_analyzer.py     # Query parsing and analysis
├── requirements.txt      # Python dependencies
├── .env                  # Environment variables (gitignored)
├── .gitignore           # Git ignore rules
├── README.md            # This file
└── screenshots/         # App screenshots

🔒 Security & Privacy

• ✅ No data stored locally
• ✅ API keys secured via environment variables
• ✅ Can be deployed in private/secure environments
• ✅ All data fetched in real-time from government sources
• ✅ No third-party data collection

🎬 Links

• 📹 Demo Video: Loom Recording
• 💻 Live Demo: Streamlit App
• 📂 GitHub Repo: priyanka7411/project-samarth

📝 Evaluation Highlights

Problem Solving & Initiative:

• Successfully discovered and integrated 2 different data.gov.in APIs
• Handled inconsistent data formats and structures
• Built end-to-end working prototype

System Architecture:

• Modular design with clear separation of concerns
• Real-time data integration from multiple sources
• Scalable and extensible architecture

Accuracy & Traceability:

• All answers include specific data points
• Every claim cites the source API
• Data fetched directly from government sources

Core Values:

• Data sovereignty: Uses free APIs, can be self-hosted
• Privacy: No data storage, secure deployment possible
• Accuracy: Real-time fetching ensures data freshness

👩‍💻 Author

Priyanka - Built for Project Samarth Challenge

�� Acknowledgments

• Ministry of Agriculture & Farmers Welfare
• India Meteorological Department (IMD)
• data.gov.in for open data APIs

---

🌾 Built for data-driven agricultural insights | Powered by data.gov.in APIs