Data Analysis for Distributed Services: Web Application for Process Discovery
Schimmel, David, Yazdi, M. Amin · Zenodo (CERN European Organization for Nuclear Research) · 2020
Process mining aims at the generation, verification, and improvement of accurate models of operational processes on the objective foundation of event logs. Various tools allow convenient applying process discovery techniques in manual process analysis. However, existing tools do not yet provide an easy way of making these techniques or the resulting process model available for automated real-time requests. Furthermore, automatically handling data coming from multiple, frequently updated data sources is complicated. These two problems make it dicult to integrate process mining applications into larger service structures in general and even more so if the structures feature numerous distributed services. In response to this challenge, the prototypical application Data Analysis for Distributed Services has been developed. It integrates the PM4Py process mining library for Python and makes process discovery techniques available as a service. It provides endpoints for all the steps, from reading data from a data source over cleaning and pre-processing that data to executing the discovery. For ease of use during manual process analysis, which still has its place, especially during the initial set-up of a discovery project, a graphical web client has been developed. Eventually, the application is evaluated on the basis of a real-world use case.