I believe most people become overwhelmed when considering the data that can be created during event processing. Number one, it is A LOT of data – and number two, the data needs real-time analysis. For the past few years, most of us have been analyzing data after we collected it, not during the event itself.
We are surrounded by technology and applications. Most every device can be connected to another device for sharing – or moving – data (except my cell phone, but it could be the operator – ha!). My favorite application is one that does real-time analytics of traffic slowdowns and street closures. My husband and I use one such app when we decide to head up the mountains on a Friday during rush hour. We check it to see where the traffic is slow, and to decide if we should use an alternate route. In this case, sensors transmit event information to a central processor for consumption by the two of us on an application on our cell phones.
If you are working on an initiative to analyze event data, don’t be overwhelmed. Instead, consider this:
- Not all data needs to be analyzed. When we analyze streaming data it is for a specific purpose. The rules are built in the processing engine to analyze and react on ONLY what we have built.
- Create questions that you need answers to using this type of processing. For example, on the production floor, how many products deviated from the specifications every hour?
- While there can be A LOT of data, it is usually short lived, and not worth keeping for years.
- Usually the results of the event processing data are combined with other data stores for later analysis.
- The rules can change over time.