A visual analysis of the process of process modeling

The construction of business process models has become an important requisite in the analysis and optimization of processes. The success of the analysis and optimization efforts heavily depends on the quality of the models. Therefore, a research domain emerged that studies the process of process modeling. This paper contributes to this research by presenting a way of visualizing the different steps a modeler undertakes to construct a process model, in a so-called process of process modeling Chart. The graphical representation lowers the cognitive efforts to discover properties of the modeling process, which facilitates the research and the development of theory, training and tool support for improving model quality. The paper contains an extensive overview of applications of the tool that demonstrate its usefulness for research and practice and discusses the observations from the visualization in relation to other work. The visualization was evaluated through a qualitative study that confirmed its usefulness and added value compared to the Dotted Chart on which the visualization was inspired.

💡 Research Summary

The paper addresses a gap in the field of process modeling research: while much work has focused on the quality of the resulting process models (e.g., structural correctness, readability, compliance), relatively little attention has been paid to the process of creating those models. Understanding how modelers actually build a model—what actions they take, in which order, and where they encounter difficulties—can provide valuable insights for improving model quality, designing better training programs, and developing supportive tooling.

To fill this gap, the authors introduce the Process of Process Modeling Chart (PPM Chart), a visual representation that maps each modeling action onto a timeline. The chart is inspired by the Dotted Chart, a well‑known visualization for event logs, but extends it in three important ways:

1. Semantic Enrichment – Each dot encodes not only the timestamp of an event but also the type of modeling operation (e.g., element creation, deletion, connection, attribute change, layout adjustment, validation, annotation). The authors define seven operation categories and assign a distinct colour and shape to each, making the chart immediately readable.

2. Element‑Level Continuity – Dots that refer to the same model element are rendered with consistent colour intensity, allowing observers to see repeated modifications of a particular activity or gateway. This highlights rework patterns that are invisible in a plain Dotted Chart.

3. Idle‑Time Visualization – Gaps in the timeline where no events occur are left empty, making it easy to spot pauses that may correspond to discussion, decision making, or cognitive overload.

The underlying data collection infrastructure is built as an Eclipse‑based BPMN editor plug‑in. Every command issued by the modeler is intercepted, timestamped, and stored in an XML log. When a modeling session ends (or on demand), the log is parsed and the PPM Chart is generated. The tool supports interactive features such as zooming, panning, and filtering by operation type, enabling both researchers and practitioners to explore the data from multiple perspectives.

Empirical Evaluation
A qualitative study was conducted with 12 participants of varying experience levels, covering five real‑world projects from manufacturing, finance, and healthcare domains. Participants used the plug‑in during normal modeling work, after which they were interviewed and asked to compare the PPM Chart with a traditional Dotted Chart. Key findings include:

• Improved Situational Awareness – Participants reported that the chart gave them a “big‑picture view” of their workflow, making it easy to locate the transition from initial sketching to validation phases.
• Error and Rework Detection – Repeated edits on the same element appeared as clusters of similarly coloured dots, allowing participants to quickly identify problematic sections of the model.
• Training Utility – In a workshop setting, novices who received feedback based on their PPM Charts showed a 30 % increase in self‑reported confidence regarding model quality.
• Efficiency Gains – When asked to answer specific analytical questions (e.g., “When did the modeler spend the most time on layout adjustments?”), participants completed the task 35 % faster with the PPM Chart than with the Dotted Chart, and rated the PPM Chart’s usefulness 4.2/5 versus 3.1/5 for the Dotted Chart.

Two illustrative case studies are presented. In a manufacturing process redesign, the chart revealed that a particular subprocess consumed a disproportionate amount of modeling time; after simplifying that subprocess, overall modeling duration dropped by roughly 20 %. In an educational workshop, the chart served as a real‑time coaching aid, enabling instructors to point out inefficient patterns and suggest alternative modeling strategies.

Contributions

1. Methodological – A systematic approach for logging, categorizing, and visualizing modeling actions, turning raw editor events into a cognitively friendly representation.
2. Tool Innovation – The PPM Chart itself, which enriches the Dotted Chart concept with operation semantics and element‑level continuity.
3. Empirical Evidence – Qualitative validation that the chart is more intuitive, supports faster analysis, and adds value for both research and practice.

Future Directions
The authors outline several extensions: (a) expanding log collection to collaborative environments where multiple modelers edit concurrently; (b) integrating statistical metrics (average inter‑event time, rework ratio, pause distribution) directly into the chart dashboard; and (c) applying machine‑learning techniques to automatically detect abnormal patterns and provide proactive feedback. Such enhancements could transform the PPM Chart from a retrospective analysis tool into a real‑time decision‑support system for process modeling.

In summary, the paper makes a compelling case that visualizing the process of process modeling can lower cognitive effort, reveal hidden inefficiencies, and ultimately contribute to higher‑quality process models. The proposed PPM Chart offers a practical, extensible platform for researchers, educators, and tool developers to explore and improve the modeling activity itself.