Real-time IoT fleet monitoring with AI-driven risk alerts — built on a Medallion Architecture using Kafka, MinIO, and Streamlit.
Global logistics teams drown in raw telemetry but lack actionable intelligence. This system converts high-frequency GPS and temperature data into proactive risk alerts — automatically flagging shipments in severe weather or exceeding thermal thresholds before damage occurs.
- Ingests real-time IoT telemetry via Kafka into a Bronze data layer (MinIO/S3)
- Enriches raw coordinates with reverse geocoding + live weather API data in parallel (Silver layer)
- Surfaces a real-time Control Tower dashboard with fleet-wide risk alerts and shipment-level tracking (Gold layer)
| Layer | Role | Tech |
|---|---|---|
| Bronze | Raw ingestion, immutable storage | Kafka, MinIO |
| Silver | Parallel enrichment (geocoding + weather) | ThreadPoolExecutor, Geopy, Weather API |
| Gold | Risk dashboard, precision tracking | Streamlit |
- Languages: Python 3.12 (Pandas, Requests, Geopy)
- Infrastructure: Kafka, MinIO, Docker Compose
- Visualization: Streamlit (real-time geospatial mapping)
- Performance: Multi-threaded parallel API enrichment via ThreadPoolExecutor
docker-compose up -d
python producer.py
python refine_with_weather.py
streamlit run app.py