Towards Understanding the Role of the Human in Event Log Extraction

Abstract

Process mining is widely used to visualize, analyze, and improve business processes. However, often its application is hindered by the considerable preparation effort that needs to be conducted by humans. One of the key tasks required in this context is obtaining the input artifact for process mining techniques: the event log. The data that is required for building such an event log typically needs to be collected from several databases and then transformed into a suitable format. While it has become clear to both academics and practitioners that the amount of human work is substantial, there is no deep understanding of the exact activities humans need to perform. Therefore, we use this paper to develop a precise understanding of how humans are involved in event log extraction. Based on a structured literature review and qualitative data coding, we derive a taxonomy of human tasks in event log extraction. This taxonomy can serve as input for both future automation efforts, as well as for process mining methodologies.