pm4py is a library for process mining in Python.
Process mining uses event logs - logs recorded during processes like moving through a healthcare pathway - to discover, monitor, and improve real-world processes. Event logs typically record “who did what, when,” allowing organizations to visualize workflows, identify bottlenecks, and verify whether operations follow intended procedures.
pm4py also supports conformance checking, comparing event logs to reference models to detect deviations, and performance analysis, revealing inefficiencies such as delays or rework loops.
pm4py’s licencing terms are quite restrictive compared to some other packages. Note that you can only use this in closed-source projects if you pay for a licence. If the code using pm4py is released open source under the AGPL licence, no costs apply at the time of writing (November 2025). This information is available at processintelligence.solutions/pm4py#licensing.
The R equivalents of this package are bupaR (which has its own buparverse including the ability to animate process mining outputs), and the R package pm4py, which provides an interface to the predictive functionality of pm4py in the R world.
Demonstration
You can try the package out in the web app embedded below (it may take a moment to load). You should not upload any real data to the app - a sample dataset is provided.
Project status
Status: Gold
Rationale: Actively maintained and widely used Python library with extensive features and user documentation.