%0 Conference Paper %1 winkler2024exploring %A Winkler, Philip %A Rosenthal, Kristina %A Strecker, Stefan %B Modellierung 2024 %C Bonn %D 2024 %E Weske, Mathias %E Michael, Judith %I Gesellschaft für Informatik e.V. %K DataModeling ModelingProcess myown %N P348 %R 10.18420/modellierung2024_004 %T Exploring Conceptual Data Modeling Processes: Insights from Clustering and Visualizing Modeling Sequences %X Research on performing conceptual data modeling finds conceptual modelers to exhibit distinct procedural patterns of data modeling: for example, when performing a data modeling task applying the Entity-Relationship Model, a repeatedly observed pattern refers to first modeling entity types, attributes and their data types, then relationship types and their cardinalities in a subsequent step. To identify patterns in data modeling processes, we cluster and visualize sequences of modeling activities of 22 conceptual data modelers at different levels of data modeling expertise. In particular, we process modeler-tool interactions in a browser-based modeling tool to visualize sequences regarding the specific modeling activity of adding entity types, attributes and relationship types to a data model, and use hierarchical clustering to identify procedural patterns based on their similarity. We find procedural patterns to follow a distinct top-down and sequential way of proceeding and identify modeling sequences with a separate phase for modeling relationship types. Our findings prepare for designing tailored modeler tool support and inform instructors and learners on the process of conceptual data modeling.

@inproceedings{winkler2024exploring, abstract = {Research on performing conceptual data modeling finds conceptual modelers to exhibit distinct procedural patterns of data modeling: for example, when performing a data modeling task applying the Entity-Relationship Model, a repeatedly observed pattern refers to first modeling entity types, attributes and their data types, then relationship types and their cardinalities in a subsequent step. To identify patterns in data modeling processes, we cluster and visualize sequences of modeling activities of 22 conceptual data modelers at different levels of data modeling expertise. In particular, we process modeler-tool interactions in a browser-based modeling tool to visualize sequences regarding the specific modeling activity of adding entity types, attributes and relationship types to a data model, and use hierarchical clustering to identify procedural patterns based on their similarity. We find procedural patterns to follow a distinct top-down and sequential way of proceeding and identify modeling sequences with a separate phase for modeling relationship types. Our findings prepare for designing tailored modeler tool support and inform instructors and learners on the process of conceptual data modeling.}, added-at = {2024-03-19T09:25:46.000+0100}, address = {Bonn}, author = {Winkler, Philip and Rosenthal, Kristina and Strecker, Stefan}, biburl = {https://www.bibsonomy.org/bibtex/284eec532e0d098237a469f3160d383ad/stefan.strecker}, booktitle = {Modellierung 2024}, doi = {10.18420/modellierung2024_004}, editor = {Weske, Mathias and Michael, Judith}, interhash = {38439795851c7936a4c5de20994412d6}, intrahash = {84eec532e0d098237a469f3160d383ad}, keywords = {DataModeling ModelingProcess myown}, note = {Potsdam, Germany. 12.-15. March 2024}, number = {P348}, publisher = {Gesellschaft für Informatik e.V.}, series = {Lecture Notes in Informatics}, timestamp = {2024-03-19T09:25:46.000+0100}, title = {Exploring Conceptual Data Modeling Processes: Insights from Clustering and Visualizing Modeling Sequences}, year = 2024 }