Process mining: what?

Beyond data mining: Problems addressable through process mining

The three main typical questions process mining can provide an answer to are:

• What has been going on in terms of system flows, until now?
• What is going on right now in terms of system flows, for real? 
• What is most probable to happen in future, in terms of system flows?

Examples of structure that is potentially better understandable via process mining include: 

• transfers in-between units of interest in a system (e.g. in healthcare: patient transfers / referrals in-between hospitals & wards; in manufacturing: pallet transfers in-between servicing stations )
• re-work by the same unit of interest in a system (e.g. in healthcare: patient readmissions, in manufacturing : pallet re-servicing by the same robotic cell)

Process mining can discover the main end-to-end highways truly followed by your processes. People often claim to follow a certain modus operandi, dictated or not by guidelines. Process discovery algorithms can shed light into what processes are really being followed, and give you a birdview model that is as close to reality as allowed by the constraints of the chosen discovery algorithm. It’s better to have a model of the process flows, with known constraints, than have absolutely no picture of what is happening.

The end-to-end performance of identified highways can be analyzed (time delays, frequencies, locations of deviations from an expected modus operandi. )

Process mining (conformance checking) can answer the question if guidelines are being followed or not.  

Process mining can bring valuable support for forecasting system flow future (in terms of time / resource allocation, probable malfunctions etc) and for overall system optimization (such as cost reduction, quality improvement, decreasing time delays).