Discovering Social Networks Instantly: Moving Process Mining Computations to the Database and Data Entry Time

Process mining aims to turn event data into insights and actions in order to improve processes. To improve process performance it is crucial to get insights into the way people work and collaborate. In this paper, we focus on discovering social networks from event data. To be able to deal with large data sets or with an environment which requires repetitive discoveries during the analysis, and still provide results instantly, we use an approach where most of the computation is moved to the database and things are precomputed at data entry time. Differently from traditional process mining where event data is stored in file-based system, we store event data in relational databases. Moreover, the database also has a role as an engine to compute the intermediate structure of social network during insertion data. By moving computation both in location (to database) and time (to recording time), the discovery of social networks in a process context becomes truly scalable. The approach has been implemented using the open source process mining toolkit ProM. The experiments reported in this paper demonstrate scalability while providing results instantly.