🌊 FloatChat

AI-Powered Conversational Interface for ARGO Ocean Data Discovery and Visualization

---

📌 Problem Statement

Oceanographic data is vast, complex, and heterogeneous. The Argo program deploys autonomous profiling floats across the world's oceans generating extensive datasets in NetCDF format. Accessing this data requires domain knowledge and technical skills — creating barriers for non-technical users.

FloatChat bridges this gap by enabling natural language access to Indian Ocean Argo data.

---

🚀 Live Demo

🔗 https://ai-float-chat.streamlit.app

---

🎯 Someexapmle that can You Ask?

"What is the average temperature in Arabian Sea in 2023?"
"Show salinity profiles near equator in January 2024"
"How many floats are in the Indian Ocean?"
"Show float locations in Arabian Sea on 01-03-2026"
"Show temperature data in Bay of Bengal in 2024"

---

🏗️ System Architecture

User Query (Natural Language)
         ↓
  Streamlit Frontend
         ↓
  Auto Fetch (if new date)
         ↓
  ChromaDB RAG Pipeline
  (Ocean knowledge context)
         ↓
  Groq LLM (LLaMA 3.3-70B)
  (NL → SQL generation)
         ↓
  PostgreSQL Database
  (3M+ ocean records)
         ↓
  Answer + Visualization
  (Maps, Graphs, Metrics)

---

🛠️ Technology Stack

Layer	Technology	Purpose
Data Source	Argo GDAC (IFREMER)	NetCDF ocean data download
Data Processing	Python, xarray, numpy	NetCDF parsing
Relational DB	PostgreSQL 18	Structured data storage
Vector DB	ChromaDB	Semantic search + RAG
Embeddings	sentence-transformers(all-MiniLM-L6-v2)	Text embeddings
LLM	Groq API (LLaMA 3.3-70B)	NL to SQL generation
Frontend	Streamlit	Chat UI + Dashboard
Visualization	Plotly	Maps + Graphs

---

📁 Project Structure

FloatChat/
├── backend/
│   ├── fetcher.py          # Real-time NetCDF download + caching
│   ├── ingest.py           # NetCDF → PostgreSQL ingestion pipeline
│   ├── sql_generator.py    # NL to SQL via Groq LLM
│   └── rag.py              # ChromaDB RAG pipeline
├── frontend/
│   └── app.py              # Streamlit chat dashboard
├── models/
│   └── vector_store/       # ChromaDB persistent storage
├── database/
│   └── schema.sql          # PostgreSQL table definitions
├── data/
│   ├── argo_cache/         # Downloaded NetCDF files (gitignored)
│   └── processed_dates.txt # Track ingested dates
├── .env                    # Environment variables (gitignored)
├── requirements.txt        # Python dependencies
└── README.md

---

🚀 Quick Start

1. Clone Repository

git clone https://github.com/YOUR_USERNAME/FloatChat.git
cd FloatChat

2. Create Virtual Environment

python -m venv venv
venv\Scripts\activate      # Windows
source venv/bin/activate   # Linux/Mac

3. Install Dependencies

pip install -r requirements.txt

4. Setup Environment Variables

Create .env file:

DB_HOST=localhost
DB_PORT=5432
DB_NAME=floatchat
DB_USER=postgres
DB_PASSWORD=your_password
GROQ_API_KEY=your_groq_api_key

5. Setup PostgreSQL Database

psql -U postgres -c "CREATE DATABASE floatchat;"
psql -U postgres -d floatchat -f database/schema.sql

6. Download Sample Data

python backend/fetcher.py

7. Ingest Data into DB

python backend/ingest.py

8. Initialize RAG System

python backend/rag.py

9. Run Application

streamlit run frontend/app.py

---

📊 Database Statistics

Metric	Value
Total Records	3,038,710+
Unique Argo Floats	265
Date Range	2021 – 2026 (Realtime)
Region	Indian Ocean (0-26°N, 55-101°E)
Data Source	Argo GDAC — IFREMER

---

🧠 NLP Pipeline

FloatChat uses NLP at 4 stages:

1. Intent Detection — Understanding what user wants
2. Entity Extraction — Location, date, parameter extraction
3. NL to SQL — Converting query to PostgreSQL SQL
4. Response Generation — Human-friendly answer from DB results

RAG Enhancement

• ChromaDB stores 15 ocean knowledge documents
• Covers Arabian Sea, Bay of Bengal, Indian Ocean facts
• Context retrieved before SQL generation for better accuracy

---

---

📈 Features

• ✅ Natural Language Chat — Ask questions in plain English
• ✅ Realtime Data — Auto-downloads latest Argo data
• ✅ Interactive Maps — Plotly geospatial float locations
• ✅ Temperature Graphs — Time series visualization
• ✅ Salinity Plots — Scatter plot visualization
• ✅ Depth Profiles — Temperature vs Depth charts
• ✅ Smart Caching — No duplicate downloads/ingestion
• ✅ RAG Pipeline — Context-aware SQL generation
• ✅ DB Stats — Live database statistics in sidebar

---

🔑 API Keys Required

Service	Purpose	Free Tier
Groq API	LLM for NL to SQL	Yes — Generous

---

📄 License

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

Built with ❤️ for Indian Ocean Science