Towards Process Mining Use Case Map Models with PM4Py-UCM

2026-06-04 · Source: cs.SE updates on arXiv.org · Field: Technology & Digital — Artificial Intelligence & Machine Learning, Software Development & Engineering, Data Science & Analytics · Depth: Expert, extended

Summary

PM4Py-UCM is an open-source extension to the PM4Py Python library, designed to integrate Use Case Map (UCM) models as a first-class output of process discovery. This tool addresses the gap between process mining and requirements engineering by enabling the automated creation and analysis of UCM models from event logs. Its key contributions include a UCM discovery pipeline, strategies for hierarchical decomposition to produce nested UCM models, configurable performer mappings for UCM and BPMN visualizations, and a bi-directional exporter to the jUCMNav tool that ensures round-trip preservation of mined models. The paper illustrates PM4Py-UCM's capabilities using public and synthetic event logs, such as an issue tracking log with 100,008 events and 9 activities, and a claims payment log with 78,126 events and 25 activities, demonstrating how behavior is rendered under different performer abstractions and decomposition strategies.

Key takeaway

For Requirements Engineers or Business Analysts seeking data-driven insights into organizational processes, PM4Py-UCM offers a direct path to generate standardized Use Case Map models from event logs. This allows you to visualize "who does what, and when" with configurable decomposition and performer mappings, directly feeding into URN-based analysis tools like jUCMNav. You should explore its hierarchical decomposition and performer binding features to gain a clearer, more actionable understanding of complex process flows and their human or system actors.

Key insights

PM4Py-UCM bridges process mining and requirements engineering by generating UCM models from event logs for enhanced analysis.

Principles

• Process models can be hierarchically decomposed.
• Performer roles can be mapped to activities.
• UCM models enable goal-process alignment.

Method

PM4Py-UCM reuses PM4Py's inductive-miner to produce a process tree, then converts it to a UCM object model, applying hierarchical decomposition and performer binding before exporting to jUCMNav.

In practice

• Use PM4Py-UCM to discover UCM models from XES/CSV logs.
• Configure decomposition for complex process visualization.
• Map log attributes (roles, resources) to UCM components.

Topics

• Process Mining
• Requirements Engineering
• Use Case Maps
• PM4Py
• jUCMNav
• Event Logs
• Process Decomposition

Code references

• ProcessMining-uOttawa/pm4py-ucm
• JUCMNAV/projetseg-update

Best for: AI Scientist, Research Scientist, Software Engineer

Related on AIssential

• The Incentive Map Every Data Platform PM Should Draw — Understanding stakeholder incentives through an "Incentive Map" enables data PMs to influence without direct authority.
• Psy-Chronicle:A Structured Pipeline for Synthesizing Long-Horizon Campus Psychological Counseling Dialogues — Long-horizon psychological counseling data can be synthesized by modeling student profiles, temporal stress events, and cross-session memory.
• Large Language Models to Enhance Business Process Modeling: Past, Present, and Future Trends — LLMs significantly advance text-to-BPMN generation but face challenges in semantic accuracy, evaluation, and real-world applicability.
• Context Graphs: AI's Next Big Idea — Context graphs capture the "why" behind enterprise decisions, enabling AI agents to understand and apply organizational precedents.
• Prototype-Grounded Concept Models for Verifiable Concept Alignment — PGCMs enhance CBM interpretability by grounding abstract concepts in verifiable visual prototypes, enabling direct inspection and intervention.
• SliceGraph: Mapping Process Isomers in Multi-Run Chain-of-Thought Reasoning — Multi-run CoT reasoning exhibits "process isomers," where diverse reasoning paths lead to identical correct answers.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by cs.SE updates on arXiv.org.