Advancing Business Process insights: A New Approach to Causal Analysis abd a Study of Event log Abstraction in Process Mining

Abstract

In modern organizations, the digital transformation has led to a rapid growth in the volume and granularity of data captured during business process execution. Operational systems continuously log events that document the activities, decisions, and interactions shaping day-to-day operations, resulting in event logs that record virtually everything that happens within a company. As this data collection continues to expand, organizations are increasingly able to analyze their processes with a level of detail and rigor that was previously unattainable, thereby unlocking new opportunities for data-driven insights and process improvement. Process mining has emerged as a prominent discipline for extracting actionable insights from event data, bridging the gap between data science and business process management. By leveraging event logs, process mining techniques aim to reconstruct actual process flows, diagnose inefficiencies, and identify opportunities for process improvement. A central challenge in this context is root cause analysis, which aims to identify the underlying factors driving undesired process outcomes. Such analyses rely on event logs that accurately and meaningfully capture process behavior. However, in many real-world settings, event data are recorded at a granularity that is too fine-grained for business users and analysts to interpret. When the level of detail in the log does not align with stakeholders’ conceptual understanding of the process, conducting reliable root cause analysis and, by extension, proposing effective improvements, becomes considerably more difficult.