The process mining field is exploring a new data model for event data called “object-centric event data” (OCED). In this data model, events are not partitioned under a unique case identifier, but each event is related to a number of data objects that can also be related to each other. The resulting data model essentially integrates classical relational data modeling with a multi-object event log defining one trace per object.
While the necessary concepts for modeling OCED are defined well, there is currently no scalable implementation for object-centric event data. For research, OCED has been implemented and validated in multiple case studies using graph databases using the labeled property graph data model.
- https://arxiv.org/abs/2410.14495
- https://github.com/multi-dimensional-process-mining/graphdb-eventlogs
- https://research.tue.nl/en/studentTheses/automated-auditing-of-controls-using-event-knowledge-graphs
- https://multiprocessmining.org/2025/07/07/5x-10x-faster-object-centric-event-data-systems-just-by-waiting-event-knowledge-graphs-are-catching-up/
However, scalability to store and execute queries for large numbers of events and objects is limited: while we have successfully executed case studies with a few million events and objects, industrial case studies require to handle 10’s or 100’s millions of events and objects.
The objective of this project is to explore and engineer options for significantly improving performance and scalability of data management of object-centric event data. Several different directions are possible, each being its own Master project:
- Using relational databases for OCED management and querying. Challenges are allowing a flexible data model to enrich data with new features and enabling high-level querying, for example by using graph query extensions over SQL databases.
- Benchmarking graph data management solutions and developing best practices for (re-)engineering data models and queries for scalable and high performance OCED data management.
- Extending and applying the graph data engine AvantGraph (https://github.com/avantlab/avantgraph) developed by the Database group of the department of Mathematics and Computer Science at TU/e to handle graph event data.
- Custom implementations of highly scalable and extensible data structured for object-centric event data.
- Other ideas from students are welcome.
Contact: Dirk Fahland (d.fahland@tue.nl)