Weather with Cloudya:
Weather with Cloudya is a full-stack web application that makes real-time climate data accessible through interactive geospatial visualizations and an AI-powered assistant. Built with a Python (FastAPI) backend and React.js frontend, it allows users to explore key weather variables—like temperature, precipitation, and humidity—on a responsive map powered by Deck.gl. The platform is fully self-hosted on a Raspberry Pi 4 using Docker containers for backend, frontend, and automated data updates. It processes and stores NOAA’s RTMA data locally, offering both historical and real-time insights. A unique feature is Cloudya, an AI chatbot powered by LLaMA 3 Vision, capable of interpreting the map visualizations and answering user questions based on the current data. The system is accessible securely via Cloudflare Tunnels, demonstrating an innovative, low-cost, and scalable approach to environmental data access. This project showcases how open data and edge computing can work together to create powerful, community-focused climate tools.
Used Technologies
- Python: FastAPI, Uvicorn, Xarray, Rioxarray, ORJSON, Herbie, SciPy, Zlib
- JavaScript: React.js, Deck.gl
- AI Integration: Meta LLaMA 3.2 Vision
- Deployment: Docker and Docker Compose
- Platform: Raspberry Pi 4 (8GB RAM), External SSD
- Cloud Services: Cloudflare Tunnel
- Version Control: Git, GitHub (SSH)
- Visualization Tools: Matplotlib, Base64 encoding