Aether-Chain

Aether-Chain is a real-time supply chain monitoring prototype built for a logistics risk prediction use case. It combines a modern frontend dashboard, a Python API that simulates live truck telemetry, and an ML model that predicts operational issues before they become critical.

This public repository is a cleaned portfolio version of the project. Cloud deployment scripts and competition-specific infrastructure details were intentionally removed from the public codebase. The focus here is the product workflow: live monitoring, predictive backend logic, and AI-assisted operations UX.

• Real-time dashboard for shipment monitoring
• Live tracking map with shipment selection
• Flask API serving simulated telemetry and aggregated metrics
• XGBoost-based risk prediction workflow
• AI chat flow for operational analysis and recommendations
• Event-driven/product-thinking mindset applied to logistics operations

• Frontend: Next.js, React, Tailwind CSS, TanStack Query, Zustand, Mapbox GL
• Backend: Python, Flask, XGBoost, pandas, NumPy
• AI integration: server-side proxy route for LLM requests

.
├── Frontend
│   ├── app
│   ├── components
│   ├── lib
│   ├── public
│   └── package.json
├── Sensor Data Api
│   ├── sensor_data_api.py
│   ├── train_model.py
│   ├── model_xgboost.json
│   ├── encoders_scaler.pkl
│   └── truck_data_version_final.csv
└── README.md

1. The Flask backend simulates truck movement, temperature, humidity, vibration, and delivery progress.
2. The backend loads a trained XGBoost model and predicts shipment risk states from incoming telemetry.
3. The frontend polls the backend and renders live shipment, alert, and tracking views.
4. The chat interface can route selected operational context to an LLM endpoint through a protected server-side API route.

cd "Sensor Data Api"
pip install -r requirements.txt
python sensor_data_api.py

The API runs on http://localhost:5000.

cd Frontend
npm install
cp .env.example .env.local
npm run dev

The app runs on http://localhost:3000.

In Frontend/.env.local:

NEXT_PUBLIC_FLASK_API_URL=http://localhost:5000
NEXT_PUBLIC_MAPBOX_TOKEN=your_mapbox_token_here
HUAWEI_LLM_TOKEN=your_llm_token_here

If you do not want to enable the AI chat integration immediately, you can still explore the dashboard and backend simulation without the LLM token.

• This repository is meant for demonstration and portfolio review.
• Sensitive cloud credentials and competition deployment assets are not included in the public version.
• The public repo focuses on architecture, product flow, and implementation quality rather than cloud account configuration.