Object-Centric Process Mining on Event Data Extracted from SAP ERP Systems

 

Object-Centric Process Mining on Event Data Extracted from SAP ERP Systems

Student: Aaron Kusters

Title: Object-Centric Process Mining on Event Data Extracted from SAP ERP Systems

Supervisor: Gyunam Park

1st Examiner: Prof. Wil M.P. van der Aalst

2nd Examiner: Prof. Dr. Stefan Decker

Summary

Enterprise Resource Planning (ERP) systems are used by organizations to manage and digitalize their business activity. Naturally, such ERP systems have the potential to be an excellent data source for process mining, which combines data science and business process management to provide insights into business processes. However, because of the complexity of ERP systems and their underlying data, general methods need to be developed to extract event logs from them and realize this high potential. Additionally, traditional process mining assumes that events correspond to a single object (i.e., a single case identifier) and therefore cannot handle the event data from ERP systems involving multiple objects very well. This thesis presents a general approach for extracting event data from the prevalent SAP ERP systems and analyzing it. For that, advancements in the process mining field, mainly object-centric approaches, are used, which are a natural fit to the data organization of ERP systems. The presented approach includes analyzing the data sources SAP ERP systems provide on business processes, how they can be retrieved, and how to utilize information in these data sources and form meaningful activities and events. Next, it presents how to develop implementations for the actual extraction of objectcentric event logs. The approach is evaluated in implementation through an accompanying production process example, for which the possible process analysis insights are explored, mimicking a case study of the production process in SAP.