Development of a method for gaining BPMN process models from real data using a process mining tool. The primary objective is the visualization and analysis of the process variants for the process of delivering packages from a single, unfiltered file with data from several thousand process instances.
Supplement
The process mining is based on event logs with scanning times that are extracted from existing systems and made available in the form of Excel lists. The open source ProM software is the process mining tool used. A converter programmed with C# first structures and sorts the data into different process groups and then converts the data into the input format XES used for ProM. With ProM, the input data are visualized as BPMN 2.0 process models and statistical evaluations (e.g. runtimes and frequencies of process variants and individual process steps). For visualization and analysis, appropriate plug-ins for ProM are identified, tested and evaluated.
Subject description
Every process step during the delivery of packages is stored in the system with an electronic time stamp (the time of scanning). With process mining, this makes it easy to analyze the chronological workflow of the process steps. Missing scans or multiple examples of the same scan lead to deviations from the expected process flow, but indicate possible weaknesses in the process (e.g. avoidable multiple sending). The knowledge gained form this serves as the basis for setting up the planned process controlling.