💬 Chat with Your Data

An intelligent AI-powered application that lets you chat with your Microsoft Fabric databases and Power BI semantic models using natural language. Built with FastAPI (Python backend) and Next.js (React frontend).

Check the video demonstration: https://youtu.be/ShrYnqdaGuc?si=SdCXLfUNGeo5Vnb_

🚀 What This App Does

• Ask Questions in Plain English: "Show me top 10 customers by revenue" or "What are sales trends this month?"
• AI Generates SQL/DAX: Automatically creates optimized queries for your data
• Self-Correcting: If a query fails, AI analyzes the error and fixes it automatically
• Real-Time Insights: Get instant answers with beautiful charts and tables
• Two Data Sources: Works with Microsoft Fabric (SQL) and Power BI (DAX)

🎯 Quick Demo

1. Connect to your Microsoft Fabric or Power BI
2. Ask "What tables do I have?"
3. Watch the AI generate and execute queries
4. Get instant results with visualizations

📋 What You Need Before Starting

Required Software

• Python 3.8 or higher (Download here)
• Node.js 18 or higher (Download here)
• Git (Download here)

Required Accounts & Services

• Azure account with admin permissions
• Microsoft Fabric workspace (for SQL data) OR Power BI Premium workspace (for semantic models)
• Anthropic Claude API key (Get one here)

🚀 Easy Setup (5 Minutes)

Step 1: Download the Code

git clone https://github.com/yourusername/chat-with-data.git
cd chat-with-data

Step 2: Run the Setup Script

# Make the script executable (Mac/Linux)
chmod +x setup.sh

# Run the automated setup
./setup.sh

Windows users: Run these commands instead:

# Copy environment files
copy backend\.env.example backend\.env
copy frontend\.env.local.example frontend\.env.local

# Setup backend
cd backend
python -m venv venv
venv\Scripts\activate
pip install -r requirements.txt
cd ..

# Setup frontend
cd frontend
npm install
cd ..

🔑 Step 3: Get Your API Keys

Get Claude AI API Key

1. Go to console.anthropic.com
2. Sign up for an account
3. Go to API Keys section
4. Create a new API key
5. Copy the key (starts with sk-ant-api03-)

Setup Azure App Registration

1. Go to portal.azure.com
2. Search for "App registrations" → Click "New registration"
3. Fill in:
   • Name: Chat with Data App
   • Account types: Single tenant
   • Redirect URI: Web → http://localhost:3000
4. Click "Register"
5. Copy these 3 values:
   • Application (client) ID
   • Directory (tenant) ID
   • Client Secret (go to "Certificates & secrets" → "New client secret")

Add API Permissions

1. In your app registration, go to "API permissions"
2. Click "Add a permission"
3. Add these permissions:
   • Power BI Service → Dataset.ReadWrite.All, Workspace.Read.All
   • Azure SQL Database → user_impersonation
4. Click "Grant admin consent"

⚙️ Step 4: Configure Environment

Edit Backend Configuration

Open backend/.env in any text editor and fill in:

# Your Claude AI API key
ANTHROPIC_API_KEY=sk-ant-api03-your-api-key-here

# Your Azure app registration details
AZURE_TENANT_ID=your-tenant-id-here
AZURE_CLIENT_ID=your-client-id-here
AZURE_CLIENT_SECRET=your-client-secret-here

# Leave these as they are
LOG_LEVEL=INFO
CORS_ORIGINS=http://localhost:3000

Edit Frontend Configuration

The frontend/.env.local file should already be correct:

NEXT_PUBLIC_API_URL=http://localhost:8000

🏢 Step 5: Add App to Your Workspaces

For Microsoft Fabric

1. Go to fabric.microsoft.com
2. Open your workspace
3. Click "Workspace settings" (gear icon)
4. Go to "Access" → "Add people or groups"
5. Enter your Application (client) ID
6. Select Member role
7. Click "Add"

For Power BI

1. Go to powerbi.microsoft.com
2. Open your Premium workspace
3. Click workspace menu (three dots) → "Settings"
4. Go to "Access" → "Add people or groups"
5. Enter your Application (client) ID
6. Select Member role
7. Click "Add"

🎉 Step 6: Start the Application

Open Two Terminal Windows

Terminal 1 - Start Backend:

cd backend

# Activate virtual environment
# Windows:
venv\Scripts\activate
# Mac/Linux:
source venv/bin/activate

# Start backend server
python main.py

You should see: ✅ Chat with Data API ready to accept requests

Terminal 2 - Start Frontend:

cd frontend

# Start frontend
npm run dev

You should see: ▲ Next.js ready on http://localhost:3000

🌐 Step 7: Test the Application

1. Open your browser and go to http://localhost:3000

2. Configure Authentication:

   • You'll see the main page
   • The authentication should show as "✅ Configured" (if not, click "Configure OAuth2" and enter your Azure details)

3. Connect to Your Data:

   • Choose either "SQL Endpoint" (for Fabric) or "Power BI Semantic Model"
   • For Fabric: Enter your workspace SQL endpoint URL and database name
   • For Power BI: Enter your XMLA endpoint URL and dataset name
   • Click "Test Connection" - you should see "✅ Connected successfully"

4. Start Chatting:

   • Try asking: "What tables are available?"
   • Try asking: "Show me 5 rows from the largest table"
   • Watch the AI generate queries and return results!

💡 Example Questions to Try

For Microsoft Fabric (SQL):

• "What tables do I have in this database?"
• "Show me the top 10 customers by revenue"
• "What are the sales trends by month this year?"
• "How many orders were placed last week?"

For Power BI (DAX):

• "What measures are available in this model?"
• "Show me sales by product category"
• "Compare this year's revenue to last year"
• "What's the total revenue for each region?"

🔍 Troubleshooting

Backend Won't Start

• Error: ModuleNotFoundError

  • Fix: Make sure virtual environment is activated and run pip install -r requirements.txt

• Error: Claude is not available

  • Fix: Check your ANTHROPIC_API_KEY in backend/.env

Frontend Won't Start

• Error: Cannot find module
  • Fix: Run npm install in the frontend directory

Authentication Issues

• Error: "Authentication not configured"
  • Fix: Check your Azure credentials in backend/.env
  • Fix: Make sure you've added the app to your workspaces

Connection Issues

• Error: "Connection failed"
  • Fix: Verify your workspace URLs are correct
  • Fix: Make sure the app has Member permissions in the workspace
  • Fix: Check if your workspace is Premium (required for Power BI XMLA)

📱 Application Features

Chat Interface

• Real-time conversation with your data
• Shows AI thinking process
• Displays query attempts and corrections
• Conversation history for follow-up questions

Data Visualization

• Auto-generated charts based on your questions
• Interactive data tables
• Export capabilities for results

Schema Explorer

• Browse your database/model structure
• View table relationships
• Explore available measures and columns

🛠️ Advanced Configuration

Performance Settings

Edit backend/.env to adjust:

MAX_QUERY_ROWS=1000        # Maximum rows returned
QUERY_TIMEOUT=30          # Query timeout in seconds
CACHE_TTL=3600           # Cache duration in seconds

Security Settings

For production deployment, generate a strong secret key:

SECRET_KEY=your-very-strong-secret-key-here

📚 Additional Resources

• API Documentation: Visit http://localhost:8000/docs when backend is running
• Detailed Setup Guide: See INSTALLATION.md
• Architecture Overview: See ARCHITECTURE.md
• Deployment Guide: See DEPLOYMENT.md

🤝 Getting Help

If you encounter issues:

1. Check the troubleshooting section above
2. Review the logs in your terminal windows
3. Verify your configuration files have the correct values
4. Test your Azure app permissions in the Azure portal
5. Open an issue on GitHub with error details

🎯 What's Next?

Once you have the basic setup working:

1. Explore Advanced Features: Try complex queries and follow-up questions
2. Connect Multiple Data Sources: Set up both Fabric and Power BI connections
3. Customize the Interface: Modify the React components to fit your needs
4. Deploy to Production: Follow the deployment guide for cloud hosting

📄 License

This project is licensed under the MIT License - see the LICENSE file for details.

---

Built with ❤️ using Anthropic Claude, Microsoft Fabric, Power BI, FastAPI, and Next.js

⭐ If this project helps you, please give it a star on GitHub!

🚀 Quick Start

1. Clone Repository

git clone https://github.com/yourusername/chat-with-data.git
cd chat-with-data

2. Automated Setup (Recommended)

# Run the setup script
chmod +x setup.sh
./setup.sh

3. Manual Setup (Alternative)

Backend Setup:

cd backend

# Create virtual environment
python -m venv venv

# Activate virtual environment
# Windows:
venv\Scripts\activate
# macOS/Linux:
source venv/bin/activate

# Install dependencies
pip install -r requirements.txt

# Create environment file
cp .env.example .env
# Edit .env with your configuration (see Configuration section)

# Start backend
python main.py

Frontend Setup:

cd frontend

# Install dependencies
npm install

# Create environment file
cp .env.local.example .env.local
# Edit .env.local with your configuration

# Start frontend
npm run dev

4. Access Application

• Open browser to http://localhost:3000
• Configure authentication
• Connect to your data source
• Start chatting with your data!

⚙️ Configuration

Backend Environment (.env)

# AI Service
ANTHROPIC_API_KEY=your_claude_api_key_here

# Azure Authentication
AZURE_TENANT_ID=your_tenant_id
AZURE_CLIENT_ID=your_client_id
AZURE_CLIENT_SECRET=your_client_secret

# Optional: Logging
LOG_LEVEL=INFO

Frontend Environment (.env.local)

NEXT_PUBLIC_API_URL=http://localhost:8000

Azure App Registration Setup

1. Go to Azure Portal → App Registrations
2. Create new registration
3. Configure API permissions:
   • Power BI Service: Dataset.ReadWrite.All
   • SQL Database: user_impersonation
4. Generate client secret
5. Note down Tenant ID, Client ID, and Client Secret

📊 Supported Data Sources

Microsoft Fabric

• SQL Endpoints: Lakehouse and Warehouse SQL endpoints
• Authentication: Azure AD OAuth2
• Query Language: T-SQL
• Features: Full schema discovery, table relationships, query optimization

Power BI Semantic Models

• XMLA Endpoints: Premium workspace semantic models
• Authentication: Azure AD OAuth2
• Query Language: DAX
• Features: Measure discovery, table relationships, model metadata

🤖 AI Capabilities

Query Generation

• Converts natural language to SQL/DAX
• Understands business terminology
• Applies proper query optimization
• Handles complex joins and aggregations

Self-Correction

• Analyzes query execution errors
• Automatically fixes syntax issues
• Retries with corrected queries
• Learns from failed attempts

Context Awareness

• Remembers conversation history
• Handles follow-up questions
• Maintains query context
• Suggests related questions

🎯 Usage Examples

Sample Questions

For Microsoft Fabric (SQL):

• "Show me the top 10 customers by revenue"
• "What are the sales trends by month this year?"
• "Compare product performance across regions"
• "Find customers who haven't ordered in 90 days"

For Power BI (DAX):

• "What are our key performance measures?"
• "Show sales by product category"
• "Compare this year's revenue to last year"
• "What's the relationship between sales and customer satisfaction?"

Advanced Features

• Follow-up Questions: "What about the previous month?"
• Filters: "Show only products with revenue > $100K"
• Comparisons: "How does this compare to last year?"
• Drill-downs: "Break that down by region"

🔧 Development

Project Structure

chat-with-data/
├── backend/
│   ├── app/
│   │   ├── auth_service.py
│   │   ├── claude_service.py
│   │   ├── enhanced_multi_agent_service.py
│   │   ├── fabric_service.py
│   │   └── semantic_model_service.py
│   ├── main.py
│   └── requirements.txt
├── frontend/
│   ├── app/
│   │   ├── components/
│   │   │   ├── EnhancedChatInterface.tsx
│   │   │   ├── FabricConnection.tsx
│   │   │   └── PowerBIMCPConnection.tsx
│   │   └── page.tsx
│   └── package.json
└── README.md

Key Components

Backend Services:

• enhanced_multi_agent_service.py: Core AI orchestration and query generation
• fabric_service.py: Microsoft Fabric database connections and operations
• semantic_model_service.py: Power BI semantic model connections and DAX execution
• claude_service.py: AI service integration and prompt management
• auth_service.py: OAuth2 authentication and token management

Frontend Components:

• EnhancedChatInterface.tsx: Main chat interface with AI thinking process
• FabricConnection.tsx: Microsoft Fabric connection configuration
• PowerBIMCPConnection.tsx: Power BI semantic model connection setup
• SchemaExplorer.tsx: Database/model structure browser

API Endpoints

Core Endpoints:

• POST /api/chat/unified: Main chat endpoint for natural language queries
• GET /api/connection/status: Check connection status
• POST /api/fabric/connect: Connect to Microsoft Fabric
• POST /api/powerbi/connect: Connect to Power BI semantic model

Authentication:

• POST /api/auth/configure: Configure OAuth2 settings
• GET /api/auth/status: Check authentication status

🐛 Troubleshooting

Common Issues

"Claude is not available"

• Check ANTHROPIC_API_KEY in backend .env
• Verify API key is valid and has sufficient credits

"Authentication not configured"

• Ensure Azure app registration is complete
• Check tenant ID, client ID, and client secret
• Verify API permissions are granted

"Connection failed"

• Verify data source URLs and credentials
• Check network connectivity
• Ensure proper licenses for Fabric/Power BI Premium

"Query generation failed"

• Check if schema was discovered correctly
• Verify AI service is responding
• Review error logs for specific issues

Debug Mode

Enable detailed logging by setting LOG_LEVEL=DEBUG in backend .env

🤝 Contributing

1. Fork the repository
2. Create feature branch (git checkout -b feature/amazing-feature)
3. Commit changes (git commit -m 'Add amazing feature')
4. Push to branch (git push origin feature/amazing-feature)
5. Open Pull Request

📄 License

This project is licensed under the MIT License - see the LICENSE file for details.

🙏 Acknowledgments

• Anthropic Claude: AI-powered query generation
• Microsoft: Fabric and Power BI platform support
• FastAPI: High-performance backend framework
• Next.js: React-based frontend framework

📞 Support

For support and questions:

• Open an issue on GitHub
• Check the troubleshooting section
• Review Azure and Microsoft Fabric documentation

---

Built with ❤️ by Sulaiman Ahmed