Process mining is a technique that uses data from an organization’s information systems to create a visual representation of some business processes. It is, by the way, a great approach to analyze and improve business processes by discovering how they are actually being executed, rather than how they are supposed to be executed.
As we’ve seen, a process is most of the time not a pure linear series of tasks. Indeed you can find in the flowchart conditions and loop which breaks the sequence. That means the analysis is different from analyzing some classical data as we have to face data splitted in a tree (contrary to a classical data approach which based on data tables). Saying in a different way you cannot analyze Process in the same way you analyze a dataset.
Nevertheless, Process Mining and Data Mining both belong to the Business Intelligence domain ! but due to their different nature they have different purposes (even if there’s obviously some overlap between these two domains).
First of all they have in common they are both driven by data and can access large volumes of information. They also use complex algorithms to discover hidden patterns and relationships into the data/processes.
| BUSINESS INTELLIGENCE | |
|---|---|
| DATA MINING | PROCESS MINING |
On the difference now:
| PROCESS MINING | DATA MINING |
|---|---|
| Focus on dynamic data (like logs) | All kind data (mainly historical data) |
| Aim to monitor real time data | Final data (no way to know more about the past of the data itself) |
| Focus on Process exceptions, issues | Analyze data patterns |
| Goal 🡪 Find out and highlight Process issues, knowledge of operations | Goal 🡪 decision guidance |
| Driven by business processes experts (Lean Six Sigma, etc.) | Driven by Data scientists |