From Relational Database to Event Log: Decisions with Quality Impact

Abstract

This paper addresses the topic of ‘Remediation approaches for event log quality assurance’. The assumption of having readily minable event logs is often not fulfilled. This paper addresses, from an end-user’s perspective, the quality issues that arise when an event log needs to be built from a relational database. The decisions that are taken when building the event log, have an impact on the quality of the event log. Namely, these decisions impact the suitability of an event log for the planned analyses. The goal of this paper is to provide an overview of the decisions that impact the quality of the event log, along with a realistic running example. Based on this overview of decisions, a procedure is presented. This procedure provides guidance to build the event log in a conscious manner, taking into account all the decisions and their impact on quality. This work relates to other studies on how to build an event log from relational databases, but puts more emphasis on how the technical decisions have a direct impact on the analyses of the practitioner that will use the event log afterwards.