FloatChat

• 

  Conversational AI for Intelligent Ocean Data Discovery and Visualization

🌊 About the Project — FloatChat

FloatChat is an AI-Powered Conversational Interface for ARGO Ocean Data Discovery and Visualization.
It enables oceanographers and data scientists to interact with complex ARGO NetCDF datasets using natural language queries, combining the power of LLMs, vector databases, and geospatial visualization tools.

🧩 Key Technical Components

• End-to-End Data Pipeline

  • Ingests and processes ARGO NetCDF oceanographic data.
  • Stores data in both:
  • Relational Database: PostgreSQL for structured metadata and time-series storage.
  • Vector Database: FAISS / Chroma for semantic search and embedding-based retrieval.

• LLM-Powered Query System

  • Backend Large Language Model (LLM) translates natural language prompts into optimized SQL and vector queries.
  • Uses Model Context Protocol (MCP) for contextual reasoning and database response synthesis.

• Interactive Visualization Dashboard

  • Frontend built with Plotly, Leaflet, or Cesium for real-time geospatial and temporal visualizations.
  • Provides tabular summaries exportable to ASCII and NetCDF formats.
  • Enables dynamic exploration of ocean parameters (temperature, salinity, depth, etc.).

⚙️ Tech Stack

• Backend: Python, FastAPI, PostgreSQL, FAISS/Chroma, LangChain / MCP
  
• Frontend: React, Plotly.js, Leaflet / Cesium
  
• Data Format: ARGO NetCDF
  
• AI / ML: OpenAI GPT / Llama Models for query translation and summarization
  

🚀 Impact

FloatChat transforms ARGO’s vast oceanographic datasets into intuitive, conversational insights, bridging the gap between AI-driven data analytics and scientific ocean data exploration.

Built With

• faiss
• fastapi
• llm
• mcp
• nextjs
• postgresql