ShockMap — Multi-Sector Supply Chain Intelligence Platform

"Know which critical inputs will run short — before headlines do."

ShockMap is a high-fidelity, three-engine intelligence platform designed to detect, propagate, and recommend responses to global supply chain disruptions. Originally focused on pharmaceuticals, it has been modernized into a sector-agnostic surface tracking Pharma APIs, Rare Earth Minerals, and Energy Security (Oil/LNG).

Built for the Google AI Hackathon 2026.

---

🛰️ Live Intelligence Flow

graph TD
    A[GDELT Global News Feed] -->|Polls every 15m| B(Engine 1: Signal Extraction)
    B -->|Gemini 2.5 Flash NER| C{Structured Shock}
    
    subgraph "The Intelligence Core"
        C --> D(Engine 2: Propagation Graph)
        D -->|Personalized PageRank| E[Downstream Risk Score]
        D -->|Louvain Clustering| F[Supply Communities]
    end
    
    subgraph "Domestic Logistics"
        E --> G(Domestic Routing API)
        G -->|Shortest Path| H[Optimal Port-to-Hospital Route]
        G -->|Bottleneck Analysis| I[Alternate Logistics Path]
    end
    
    subgraph "Action Intelligence"
        E --> J(Engine 3: RAG Analyst)
        J -->|Qdrant Knowledge Retrieval| K[Grounded Action Plan]
    end
    
    K --> L((React Intelligence Surface))
    H --> L
    F --> L

Loading

---

🏗️ 3-Engine Architecture

Engine 1 — Signal Ingestion & Extraction

• Polling: Scrapes GDELT every 15 minutes for 64 supply-critical keywords.
• NER (Gemini 2.5 Flash): Extracts structured data from raw news: province, severity, event_type (shutdown, export_ban, etc.).
• Sector Filtering: Automatically routes signals to Pharma, Rare Earth, or Energy pipelines.

Engine 2 — Graph Propagation & Clustering

• Graph Topology: 20 Pharma APIs, 8 Rare Earths, 16 Energy nodes, 34 Source Provinces.
• Mathematics: Runs Personalized PageRank to model how a localized shock (e.g., Hebei factory fire) cascades into national drug shortages.
• Community Detection: Uses Louvain to surface hidden supply dependencies between uncorrelated sectors.

Engine 3 — RAG Action Intelligence

• Knowledge Base: 2,400+ entries across NLEM, SARS-2003 playbooks, CDSCO policies, and custom logistics benchmarks.
• Retrieval: Uses Gemini Embeddings + Qdrant Cloud.
• Output: Generates specific, grounded procurement advice — including supplier alternatives and emergency lead times.

---

🧪 Geopolitical Scenarios & Domestic Routing

⚓ Hormuz Crisis Tracker

Simulates a 59-day blockade of the Strait of Hormuz.

• Impact: Brent Crude spike to $118/bbl, tanker diversion (+18 days), and 34% increase in pharma cold-chain costs.
• Refinery Watch: Real-time utilization tracking (e.g., Reliance Jamnagar at 64% capacity).

🚛 India In-Depth (Domestic Supply Chain)

Full internal logistics visibility for critical APIs:

• Node Flow: Port → CWC Warehouse → Formulation Plant (Cipla, Dr Reddy's, etc.) → Regional Distributor → Hospital.
• Dynamic Routing: Automatic rerouting around bottlenecks like JNPT congestion or Baddi Pharma Hub stockouts.

---

🛠️ Tech Stack

• Backend: Python 3.11, FastAPI, Uvicorn.
• AI/ML: Google GenAI (Gemini 2.5 Flash), NetworkX (Graph Math).
• Vector DB: Qdrant Cloud (RAG Storage).
• Frontend: React 19, Vite, Recharts (Data Viz), CesiumJS (3D Globe).
• Maps: ISRO Bhuvan (WMS), OlaMaps, MapMyIndia (Mappls).

---

🚀 Running Locally

1. Prerequisites

• Gemini API Key (AI Studio)
• Qdrant Cloud Credentials

2. Backend Setup

cd backend
pip install -r requirements.txt
# Set Env Vars
$env:GEMINI_API_KEY="your-key"
$env:QDRANT_URL="your-qdrant-url"
$env:QDRANT_API_KEY="your-qdrant-key"
python -m uvicorn app.main:app --host 127.0.0.1 --port 8000 --reload

3. Frontend Setup

cd frontend
npm install
npm run dev

Navigate to http://localhost:5173.

---

🔮 Roadmap & Future Upgrades

• [ ] GNN Propagation (v3.0): Transition from PageRank to Graph Neural Networks (GNN) for predictive temporal risk modeling.
• [ ] Multi-Agent Procurement: Deploy Gemini agents to automatically negotiate "emergency reservations" with alternate suppliers via API.
• [ ] Dynamic HHI: Live Herfindahl-Hirschman Index monitoring for every Indian state's import concentration.
• [ ] Digital Twin Integration: Sync with real-time port telemetry (AIS data) for minute-by-minute container tracking.

---

📜 License & Credits

Built by ShockMap Intelligence Team for the Google AI Hackathon. License: MIT.

ShockMap demonstrates that Palantir-grade supply intelligence is achievable using open signals, graph mathematics, and grounded LLM reasoning.