LoVizQL: A Query Language for Visualizing and Analyzing Business Processes from Event Logs

Process event logs record information about the execution of the activities of a business process. Process mining techniques use these event logs to discover, analyze, and optimize business processes. Process mining tools offer many functionalities such as data filtering, process discovery, process visualization, or conformance checking. Process visualization is generally based on Directly-Follows Graphs (DFGs), where each node represents an activity of the process, and each transition represents a directly-follows relation between nodes (activities). A workflow frequently followed by process mining analysts involves manually comparing the DFGs of different event log subsets (e.g., subsets belonging to different product categories in a purchase-to-pay process) to identify patterns or behaviors in the process data (e.g., delays in process execution). However, performing this type of analysis with current process mining tools is usually a time-consuming task, especially if the number of event log subsets analyzed is large. This research aims to address this limitation by presenting LoVizQL, a query language to obtain collections of DFGs that meet specific user-defined conditions in the queries. The language is evaluated using reports belonging to various Business Process Intelligence Challenges. The evaluation demonstrates that LoVizQL covers analyses found in real scenarios and reduces the effort to find specific subsets of event log data and their corresponding DFGs.