IoT monitoring dashboard for industrial sensors

§ 01Brief

The plant was running six independent SCADA systems with no unified view. Operators had to watch six separate screens and manually correlate anomalies across production lines. Alerts were email-based and frequently missed.

We replaced this with a single React dashboard consuming a WebSocket feed from a new Node.js aggregation layer. Alerts became in-app with sound and visual escalation. The ops team went from six screens to one.

§ 02Approach

1. 01
   Architecture

   WebSocket aggregation layer

   A Node.js service polled all six SCADA APIs at 500ms intervals, normalised the data schema, and pushed diffs to clients over a single WebSocket connection. Client payload dropped by 80% vs polling.

2. 02
   Design

   High-density data layout

   1,200 sensors across 6 lines — we ran a card-sorting session with 4 operators to build the information hierarchy. The result: a zone → line → sensor drill-down with persistent anomaly indicators.

3. 03
   Alerts

   Tiered alert system

   Three severity tiers — info, warning, critical — each with distinct colour, sound, and escalation behaviour. Critical alerts locked the UI into full-screen mode until acknowledged.

4. 04
   Performance

   Sub-second render at full load

   React virtualisation for the sensor grid. Canvas-rendered sparklines for historical trends. The dashboard held 60fps at full 1,200-sensor load on the plant's standard-issue hardware.

§ 03Outcomes

§ 04Gallery