The client needed a single platform to make sense of data flowing from a large fleet of connected devices — not just to store it, but to visualise it live, analyse it, and route it on to the systems that act on it.
The Brief
IoT data arrives fast, in high volume, and in inconsistent formats. The platform had to reliably ingest that stream, normalise it through data parsers, surface it in real-time dashboards, and forward clean, structured data to downstream consumers — all without dropping messages under load.
What We Built
- A high-throughput ingestion layer built on VerneMQ, handling device messaging over MQTT at scale
- Modern, modular data parsers that normalise incoming device payloads into a consistent schema
- Real-time dashboards for visualising live device data as it streams in
- An analytics layer for querying, aggregating and deriving insight from both live and historical data
- A forwarding pipeline that routes parsed data on to downstream services and integrations
- A dual-database design — MongoDB for high-volume device telemetry and PostgreSQL for structured, relational data
Built on a Real-Time Stack
The platform is built with React on the front end and Node.js on the back end, with VerneMQ as the MQTT broker at the ingestion edge. MongoDB stores high-volume, schema-flexible device telemetry while PostgreSQL handles structured relational data — giving the platform both the throughput to keep up with the devices and the query power to analyse what they send.
The result is a single platform that takes raw device data and turns it into something visible, analysable and ready to act on — in real time.