Computer Science · intermediate
Distributed Edge IoT Monitoring System
A system that collects simulated IoT sensor data, processes it at the edge via Docker, and sends only critical alerts to a dashboard.
Project overview
As cloud costs rise, "Edge Intelligence" is becoming a massive field. This project shows you understand the hardware-software interface and how to build efficient, low-latency systems.
Real-world use case
A manufacturing plant monitoring machine health to predict failures before they happen, saving millions in downtime.
Tools and technologies
- Python
- Docker
- MQTT
- InfluxDB
- TensorFlow Lite
Skills you can demonstrate
Distributed Systems, Edge Computing, MQTT
Build roadmap
- Design the sensor simulators
Python scripts that mimic vibration, temperature, and pressure sensors publishing to MQTT at realistic rates, with injected anomalies you can trigger on demand. - Set up the MQTT broker
Run Mosquitto in Docker and confirm messages flow with MQTT Explorer. Define a topic hierarchy per site and sensor so the edge processor can subscribe by wildcard. - Build the edge processor
A container that subscribes to all sensor topics, runs a TensorFlow Lite anomaly model on a sliding window, and forwards only critical events. Measure its memory footprint. - Store time-series data
Add InfluxDB to the Compose file. Write normal readings to a bucket with short retention and alerts to one kept longer. Query both to prove the split works. - Build the Grafana dashboard
Live sensor panels, anomaly rate over time, and a health row showing broker and processor status. Trigger an injected anomaly and watch it arrive on the dashboard. - Package with Docker Compose
One command starts the broker, three sensors, the edge processor, InfluxDB, and Grafana. Record a clean start from nothing to a live dashboard for the README.
Ways to extend the project
Use a local LLM (via Ollama) at the edge to summarize sensor logs into human-readable "Status Reports" every hour.
How to showcase your work
- GitHub: A docker-compose.yml that spins up the entire distributed environment with one command.
- LinkedIn: Screenshot of your Grafana dashboard showing real-time anomaly detection.
- Resume: Focus on edge computing, distributed systems, low-latency inference.