What is process mining?

[INTRO MUSIC PLAYING] Hi, welcome back. In this lecture, I would like to explain to you what process mining is. We start from the observation that event data is recorded everywhere. For instance, when you use your bank card to buy groceries, or whenever you make a phone call, or send an email, event data is recorded. Similarly, when you use your smart TV and change channels, or your smartphone and open or close apps, event data is recorded. Also when you use your Oyster, or public transportation card, event data is created. Similarly, in a hospital or health care environment, event data is recorded and stored, and this can be used for further analysis. So what do we consider a process?

So let’s look at a very simple example of a patient treatment. So when a patient comes in, there can be a first consult, where the doctor investigates what it could be. In this example, the doctor prescribes three further diagnoses to be handled or executed in parallel, but all three have to be executed. The first test that she wants done is an x-ray scan. Secondly, a blood test needs to be performed, and thirdly, a physical test needs to be performed to see what the current status is. Once these three tests have been performed, a second consult will be planned where the doctor consults with the patient to see what the disease might be.

Based on this second consult, a treatment is prescribed. Either surgery is necessary, or it can be cured through medicine. And after some period, there’s a final consult, where the doctor and the patient meet for the final time, and the treatment of this patient is considered done. So this is a very simple process. But I hope you can see there are several points where data could be recorded and left behind in the information system. So I’m sure that you know several software systems that you use in your daily practice, or that are used in your health care organisation, that supports processes. So also in a health care environment, these software systems exist, and they are there

to support the real world: Patients coming in, doctors performing exams, medical scans, surgery, etc, etc. Also bills need to be paid, salaries need to be paid, etc, etc, all supported by software systems. So the software system tries to control, in a way, the world. This software system, then, is configured using a process model. This process model describes the real world and what is expected, and therefore, is used to configure the software system. to appropriately react on incoming triggers. Once this software system is used by the real world, event data, or more particularly, event logs, are created. So this event data leaves a trace of what happened in the real world.

And process mining tries to bridge the gap between the models or the process models that were used to configure the system, and the event data that was recorded in the system. And there are three key elements of process mining. Process discovery, which solely on the base of the event data, tries to discover a process model to describe what really happened. Secondly, conformance checking. So given a process model that was used to configure the system and the event data describing what really happened, overlay them on each other to see where one of the two deviated from the other. And thirdly, enhancement.

So given a process model, either discovered or already pre-made, and the data, the process model can be overlaid with performance information to show where bottlenecks are, for instance. So in this course, we will discover all three aspects of process mining, specifically for the health care setting. So I hope you enjoyed this lecture and hope to see you in the next one.