Descriptive analytics – insight into device health

Descriptive analytics, or threshold analytics, does basic computation based on past data to generate some meaningful insights. These analytics consist of very simple computation logic, applied typically over a time range of data from the device's sensors. Parameters for the thresholds are typically given by the device manufacturers, hence identifying any deviation from the threshold can provide valuable insights into the health of the device. Descriptive analytics is useful because it allows us to learn from past data, and understand the health of the machine. In a typical real life scenario some of this data could come from sensors and others could be collected over time from manual readings. Analytics could be triggered based on firing of an event, or scheduled at the given time frame.

The vast majority of the IIoT analytics fall into this category. These analytics can also be used to detect any maintenance events, such as when to power wash a turbine or when to replace a part. We can use basic technology such as scheduler to execute these programs, and various programming languages to implement these types of analytics.