Sentinel v3.0: AI-Powered Supply Chain Intelligence

"Predict the Risk, Command the Solution."

---

💡 Inspiration

Global trade is highly vulnerable; a single disruption can cause a multi-million dollar butterfly effect. Traditional dashboards only show what happened, but they rarely tell you what to do next. I built Sentinel v3.0 to bridge the gap between static visual analytics and strategic action, transforming a Tableau Geospatial Intelligence map into a proactive AI Agent.

🚀 Key Features & The Tableau Role

Unlike static reports, Sentinel uses Tableau as its primary interactive engine and "eyes":

• Visual Command Center: Real-time tracking of over 240 vessels via an interactive Tableau Dashboard.
• Event-Driven AI: The AI analysis is triggered directly by user interactions within Tableau using the Tableau Embedding API v3 .
• Live Intelligence: Enriches Tableau's shipment data with live "ground-truth" weather conditions (wind, temperature, pressure) via OpenWeather API.
• Strategic AI Recommendations: Google Gemini 3 Flash processes Tableau's metadata to provide executive-level intervention plans.
• Actionable Alerting: Results are pushed to Slack with interactive "Block Kit" buttons for immediate approval and execution.

🛠️ Built With

Sentinel v3.0 integrates a modern tech stack to create a high-performance AI Agent:

• Languages: Python 3.x (Backend), JavaScript (Frontend), HTML5, CSS3.
• Backend Framework: FastAPI (Asynchronous, High-Performance) with Uvicorn.
• Artificial Intelligence: Google Gemini 3 Flash (Generative AI).
• Visualization & Trigger: Tableau Cloud, Tableau Embedding API v3.
• External APIs: OpenWeather API (Maritime Intelligence) and Slack Webhooks (Crisis Communication).

📊 Financial Exposure Model

Sentinel quantifies risk using a dynamic financial model to prioritize high-value interventions:

$$E_{total} = V_{cargo} + (C_{hourly} \times T_{delay})$$

• $E_{total}$: Total Financial Exposure.
• $V_{cargo}$: Cargo Valuation (Retrieved from Tableau metadata).
• $C_{hourly}$: Penalty Cost per hour of delay.
• $T_{delay}$: Estimated Duration of the disruption.

🏗️ Technical Architecture

The following diagram illustrates how Tableau acts as the frontend trigger for the entire AI ecosystem:

graph TD
    A[Tableau Dashboard] -- "Mark Selection (Embedding API v3)" --> B(Frontend Control Panel)
    B -- "Shipment Data (ID, Location, Reason)" --> C{FastAPI Backend}
    C -- "Request Live Weather" --> D[OpenWeather API]
    D -- "Ground-Truth Data" --> C
    C -- "Strategic Prompt" --> E[Google Gemini 3 Flash]
    E -- "Executive Analysis" --> C
    C -- "Rich Format Report" --> F[Slack Channel]
    C -- "Update UI" --> B
    F -- "Action Buttons (Approve/Reroute)" --> G[Operational Execution]

Loading

⚙️ Installation & Setup

1. Clone the Repository:

   git clone [https://github.com/kerememree/sentinel-supply-chain-ai.git](https://github.com/kerememree/sentinel-supply-chain-ai.git)
   cd sentinel-supply-chain-ai

2. Install Dependencies:

   pip install fastapi uvicorn requests google-generativeai python-dotenv

3. Configure API Keys: Create a .env file and add your keys:
   • GEMINI_API_KEY
   • OPENWEATHER_API_KEY
   • SLACK_WEBHOOK_URL
4. Run the Server:

   python server.py

---

Powered by Sentinel v3.0 | Weather: OpenWeather API