MOSDAC Help-Bot: Next-Gen Geospatial Intelligence

Prototype

🎥 Demo Walkthrough

Unlock the Future of Satellite Data with AI-Powered Insight

We present MOSDAC Geospatial Help-Bot, a smart AI assistant designed to simplify access to meteorological and satellite data on the MOSDAC portal. The bot supports natural language queries, processes both static and dynamic content, and answers user questions with source-linked context and geospatial coverage maps. It combines a Retrieval-Augmented Generation (RAG) pipeline (using FAISS + BM25), a knowledge graph, and LLMs like GPT or DeepSeek to deliver accurate, explainable responses. Features include entity recognition, region-aware search, product lookup, and a clean Streamlit interface with performance metrics. The bot is fully modular, resource-efficient, and adaptable for other ISRO web platforms.

---

Key Features

• 🛰️ Satellite Mission Intelligence

  • Deep-dive into Oceansat-2, ScatSat-1, INSAT-3D/3DR and more
  • Instant access to mission specs, data products, and live status

• 🌊 Oceanographic Data Fusion

  • Real-time sea surface temperature & ocean color analytics
  • Dynamic wind vector & ocean current visualizations

• 🌀 AI-Powered Cyclone Tracking

  • Predictive cyclone modeling and impact forecasting
  • Interactive historical cyclone analysis and live updates

• 📊 Hyper-Contextual Geospatial Intelligence

  • Location-aware data coverage and smart product suggestions
  • Advanced spatial analytics and scenario simulation

• 🔍 Autonomous Knowledge Graph

  • Automated knowledge extraction and live entity visualization
  • Natural language & semantic search across the knowledge base

• 📈 Self-Monitoring Performance Metrics

  • AI learning: intent accuracy, response completeness & consistency
  • Transparent analytics dashboard for continuous evolution

---

Getting Started

# Clone the future-ready repository
git clone https://github.com/AkshitTiwarii/MOSDAC.git
cd MOSDAC

# Install next-gen dependencies
pip install -r requirements.txt

# To run locally: python single_app.py

# Launch the AI dashboard
streamlit run single_app.py

---

Experience the Future

• Ask natural, futuristic queries:

  “Which satellites are monitoring the Indian Ocean right now?”
  “Simulate cyclone impact for 2030 scenarios.”
  “Visualize the knowledge graph for ocean salinity data.”

• Explore Real-Time Dashboards:

  • Monitor live satellite status and mission timelines
  • Interact with dynamically updating knowledge graphs
  • Track AI performance with live metrics
  • Tested with 50 queries p, avg latency 1.7s, intent accuracy 92%

• Continuous Evolution:

  • Auto-refresh knowledge base from the latest data streams
  • Modular architecture for rapid upgrades
  • Seamless integration with future MOSDAC missions
  • Fallback to local LLM(Deepseek/Ollama/GPT-4) in case API not allowed

---

System Architecture

MOSDAC Help-Bot
├── Automated Knowledge Engine
│   ├── Web Crawler 2.0
│   ├── Content Extractor (HTML/PDF/Live Feeds)
│   └── Knowledge Graph AI Builder
│
├── Adaptive AI Core
│   ├── Neural Retrieval System (BM25+)
│   ├── Deep Intent Recognition
│   ├── Entity Extraction AI
│   └── Contextual Response Generator
│
├── Self-Evaluating System
│   ├── Automated Intent Accuracy Tracking
│   ├── Entity & Response Metrics
│   └── Consistency Self-Monitoring
│
└── Immersive User Interface
    ├── Space-Infused Dashboard
    ├── Real-time Conversational AI
    ├── Live Satellite Mission Panel
    └── Interactive Analytics

---

Screenshots

---

The MOSDAC Help-Bot is an advanced AI assistant built to transform the way you access, analyze, and understand meteorological and oceanographic satellite data from the Meteorological & Oceanographic Satellite Data Archival Centre (MOSDAC). Experience seamless, intelligent geospatial discovery and real-time insights with an interface designed for the explorers of tomorrow.

---