AI-Powered Conversational System for ARGO Float Data (MVP)

Project Structure

• data_ingestion/ — Scripts to parse ARGO NetCDF and load into PostgreSQL/Parquet
• backend/ — RAG pipeline, LLM integration (Ollama/HF), Model Context Protocol
• frontend/ — Streamlit dashboard, chatbot, geospatial visualizations
• vector_db/ — FAISS/Chroma setup and metadata scripts
• requirements.txt — Python dependencies
• README.md — Project overview and setup instructions

Features

• Ingest ARGO NetCDF, convert to SQL/Parquet
• Store metadata in FAISS/Chroma
• RAG backend with LLM (Ollama/HF)
• Streamlit dashboard with maps, plots, chatbot
• Demo with Indian Ocean ARGO data
• Extensible to other ocean datasets

Quickstart

1. Install dependencies: pip install -r requirements.txt
2. Run data ingestion: python data_ingestion/ingest_argo.py
3. Start backend: python backend/server.py
4. Launch dashboard: streamlit run frontend/app.py

Data Sources

• NOAA ARGO repository

Extensibility

• Add new data sources by extending data_ingestion/
• Add new LLMs or RAG logic in backend/
• Add new visualizations in frontend/