In the Event Analyzer, the new extension for CloverETL Designer, the ability to process time-based data in the CloverETL environment is simple – and incredibly useful.
With the Event Analyzer, you can now process and analyze data with time-based characteristics, including log records, transactions, measurements, alerts, and more, all in the CloverETL environment. This new addition, available in beta, has many uses. As time is a key attribute in our world, data in relation to it can lead to valuable insights for businesses.
What’s the Benefit?
Sometimes you can’t know what you’re looking for until you’ve found it. With the Event Analyzer, uncover inconsistencies and truths in data to help you rework and rewire commercial processes – or even just understand your customers and the daily activities in your business better.
The Event Analyzer gives users a look into customer actions, fraud or unusual behavior, inefficient systems, discrepancies between systems, and even SLA violations – in essence, valuable information hidden in the sea of data and commotion. The extension can help you to understand your time data better so you can put this information to good use.
The Event Analyzer provides a powerful set of components to process records in the context of time such as:
- FollowRecogniser for detection of event sequences
- NonFollowRecogniser for missing events
- ChangeRecogniser for changes in the flow of events
- And RunningAggregate for computing time-based characteristics
A Real-World Example
The video below details an analysis of records from an online retailer. In this example, the e-shop analyst wants to view and understand two things involving the purchasing customer. The first, when a customer first enters the store, and secondly, when he or she made a purchase. An additional stipulation is a defined time frame: the analyst only wants the users who purchased something in ten minutes or less. With the Event Analyzer, these specifications are easy to set.
How is this done?
We can see the records of when a visitor came to the e-shop and also when he or she made the order. We sort these records chronologically. Next, we correlate the two time events and split the data into two event streams for further analysis: the first stream being the entrance and the second being the purchasing event, or “order confirmation.”
To look for an event sequence where the landing page visit is followed by the order confirmation page in less than 10 minutes, we use the component: FollowRecognizerTwoPorts. It’s important to configure the component by defining the file with an event occurrence timestamp for both input ports. Also, setting the parameter “Join key” to value “user session” ensures that there will be events connected only with the same user session.
Detection results are sent to the output port of the component and the component allows users to transform the output data using CloverETL transformations. The output of the graph is the filtered result set based on the rules defined in the components.
And with that, the e-shop analyst now can clearly see all customers who purchased within ten minutes – insight he did not have before using the EA extension. Make sure to watch the whole video for the step-by-step view of this example.
Easy as 1, 2, 3
As you can see, there’s much you can learn from your time-based data. A deeper look can lead to new ideas, changes, improved security measures, and renewed value for your business. Take advantage of the possibilities. Learn more about what’s going on with the Event Analyzer Extension for CloverETL. Download it in beta today.