Visualizing business process evolution

Abstract

Literature in business process research has recognized that process execution adjusts dynamically to the environment, both intentionally and unintentionally. This dynamic change of frequently followed actions is called process drift. Existing process drift approaches focus to a great extent on drift point detection, i.e., on points in time when a process execution changes significantly. What is largely neglected by process drift approaches is the identification of temporal dynamics of different clusters of process execution, how they interrelate, and how they change in dominance over time. In this paper, we introduce process evolution analysis (PEA) as a technique that aims to support the exploration of process cluster interrelations over time. This approach builds on and synthesizes existing approaches from the process drift, trace clustering, and process visualization literature. Based on the process evolution analysis, we visualize the interrelation of trace clusters over time for descriptive and prescriptive purposes.