Use case and data
We have four artificial sensors installed in the field and our digital control system collects the information from these sensors. We need to put a real-time analysis system in place. We need a feature in that system to cluster the incoming data and display those clusters in real time in a digital dashboard. The idea is users watching those dashboards will be alerted immediately for some anomaly or malfunction if they see unusual cluster patterns.
This example is inspired by the sensor network example given in pubnub at: https://www.pubnub.com/developers/realtime-data-streams/sensor-network/.
The following diagram should sufficiently visualize our use case:
What we have here is a lambda architecture (http://lambda-architecture.net/).
Our digital control system is generating data at a particular rate. It has the following sensors installed:
- Radiation sensor (millrads/hour)--the range of data is between 195 and 202
- Humidity (%)--the range of data is between 74 and 82
- Temperature...
