Live industrial data: OPC UA, MQTT and CSV feeds in InduQT
Real public data streaming into InduQT right now: a boiler on a public OPC UA server, Helsinki trams over MQTT and a NOAA tide gauge over a CSV link. The same connectors your plant uses, and the same decision engine watching them.
Connecting to the live feeds…
What InduQT does with a feed
Every signal gets a name, units and a normal range. Every two minutes the decision engine checks machines with fresh data; when a signal leaves its range it profiles the last two hours, finds which signal moved first and in which direction, ranks the causes it knows for that kind of equipment, and opens an investigation with the check that separates them. The same engine found three of four seeded faults in a public DOE boiler-plant data set; see how it was tested.
Questions
Where does this data come from?
From public sources: the OPC UA sample server run by OPC Labs at opcua.demo-this.com (a simulated boiler with drum level, feedwater and steam flow), Helsinki Regional Transport’s open high-frequency positioning feed (mqtt.hsl.fi, real trams: speed, acceleration, schedule delay, doors) and NOAA Tides & Currents station 9414290 at San Francisco (real water level every 6 minutes).
Is it the same as connecting a plant?
Yes. Each feed uses an InduQT connector you can set up yourself: OPC UA (node ids mapped to signals), MQTT (topic filters, a JSON field per signal) and a scheduled CSV link. Node-RED, historians and proprietary systems post to the ingest API with a site key instead.
Can Claude read it?
Yes. With the InduQT connector in Claude, ask “show me the public demo machines” and Claude reads these feeds, or connect your own OPC UA server or MQTT broker from the conversation.
Data: HSL open data (CC BY 4.0, Helsinki Region Transport); NOAA CO-OPS (public domain); OPC UA sample server by OPC Labs (simulated). Feeds can pause when a source is offline.