Master Thesis - Process Mining Case Studies on SAP


Master’s Thesis – Process Mining Case Studies on SAP

Important note: the thesis is realistically feasible only if the student has already an ongoing collaboration/work contract with a company actively interested in analyzing SAP data without relying on commercial process mining software (Signavio, Celonis, ...)

Process mining has gained traction in recent years in analyzing real-life business processes, including customer relationship management (CRM) and enterprise resource planning (ERP) systems. Some commercial vendors of process mining tools have a large set of customers and successful case studies in which the application of process mining techniques reduced the operational costs and increased customer satisfaction.

The Process and Data Science group of the RWTH Aachen University offers excellent students the possibility to write an MSc thesis on applying process mining techniques in the context of a real-life SAP ERP system.

The candidate student can propose the partner/context of the project. In particular, different processes supported by SAP (for example, the Order-to-Cash, Procure-to-Pay or Accounts Payable/Receivable), or different process mining techniques can be the target of the thesis:

  • Graph-related techniques:
    • Navigation of the document flow of different SAP processes:
      • Process querying.
      • Identification of meaningful components in the graph.
      • Clustering / concept drift detection techniques.
    • Graph similarity techniques (identifying commonalities between different executions).
  • Object-centric process mining techniques (process discovery and conformance checking): events in SAP are often related to different objects. Hence, object-centric process mining can provide meaningful insights.
  • Conformance / Compliance checking techniques:
    • Anomaly Detection (i.e., identifying document flows that are significantly different, for example, payments not preceded by an invoice).
    • Fraud Detection (different types of fraud exist; for example, duplicated payments, or payments always slightly under the amount that requires approval).
  • Performance analysis:
    • Identification of the “happy paths” inside the process.
    • Calculation of the cost of inefficient executions.
  • Confidentiality/privacy techniques:
    • Anonymous event data collection and publishing.
    • Identifying and anonymizing confidential information.
    • Developing data access control techniques.
  • Automation techniques:
    • Recommendation systems.
    • Simulation of processes in SAP.

The output of the project should be preferably a real-life case study on the application of some process mining techniques on top of the SAP ERP system of the partner. In alternative, a theoretical thesis on the development of one of the aforementioned areas is also possible.

The output should include a report (with a maximum of 80 pages) and the code implemented by the student; both released with an open-source license (for example, MIT/BSD/Apache 2.0 for the code, and CC 4.0 for the thesis). The report should not contain business-critical information on the partner. In addition to that, a report and a presentation of the results to the partner is required. The usage of different programming languages (Java, Python, …) is possible.