Bringing in datasets from technician repair visits, maintenance records, and call center logs is another great place to start tapping value from your data. There will be key events such as repair work or complaint calls that can be linked to unusual patterns in IoT data.

Maintenance records will be useful in developing predictive maintenance models. It may also tie into equipment failure predictive models. Call center logs could be mined to develop attributes of IoT data events beyond just reported numericals or categorical values. Natural Language Processing (NLP) can be used to extract attributes from text data.

For a utility company that provides natural gas, for example, the phrase rotten egg in call log texts, corresponding to a pattern in the IoT data for low gas pressure, could be a tip off that a gas leak is occurring. Analytics could then be developed to recognize the pattern in future IoT data records for proactive investigation.