Wynn Resort Forecast AI

Real-time hotel operations forecasting using AI agent and WebSocket communication.

Quick Demo

1. Set your Anthropic API key in .env file:

   ANTHROPIC_API_KEY=your-key-here
   

2. Run the application:

   docker-compose up --build

3. Open browser:

   • Frontend: http://localhost:3567

4. Try these commands:

   • "Big UFC fight this Saturday"
   • "Convention next Monday morning"
   • "Pool party season starting"

Project Structure

AgentForecast-Hotel/
│   .env                    # Your Anthropic API key
│   .gitignore
│   LICENSE
│   README.md              # This file
│   docker-compose.yml     # Runs everything
│
├───backend/
│       agent.py           # Anthropic AI agent (ForecastAgent)
│       data.py            # Data generation utilities
│       Dockerfile         # Backend container
│       main.py            # WebSocket server (glue code)
│       model.py           # RNN forecast model (ForecastModel)
│       requirements.txt   # Python dependencies
│
└───frontend/
        app.jsx            # React application
        Dockerfile         # Frontend container
        index.html         # Main HTML
        styles.css         # Styling

How It Works

1. RNN Model generates a 7-day hourly forecast at startup
2. WebSocket connects frontend to backend in real-time
3. AI Agent interprets natural language and modifies forecast
4. Live Updates broadcast changes to all connected clients

Architecture

• Backend: Python WebSocket server on port 8567
• Frontend: React app served by nginx on port 3567
• AI: Anthropic Claude for natural language understanding
• State: Single shared forecast modified by agent

Troubleshooting

• Connection failed? Check if Docker is running
• Port conflict? Change ports in docker-compose.yml
• API key error? Verify .env file has correct key
• Import errors? All backend files should be in flat structure

License

MIT