KPI-Driven Visualizations for Root Cause Analysis in Manufacturing Processes
摘要
In the past years, process mining has been a key technology to improve business processes. For example, through outlier, often also referred to as anomalies, detection and root cause analysis. However, visualizations for root cause analysis and anomaly detection have often been missing. Hardening it to follow up on anomalies through adaptive decision-making. Such visualizations aim at helping manufacturing professionals analyze this data on their own - without a need for a data expert. Existing works exploring this topic do not focus on the manufacturing domain, thereby missing the unique complexities inherent to this sector. We bridge this gap by introducing an interactive KPI-driven dashboard with two views: one for the overall material overview and one for the detailed analysis. This dashboard provides understandable KPI-driven visualizations for detecting anomalies and supports the identification of their root causes. Enabling the identification and assessment of unplanned, ad hoc, production approaches. The dashboard’s benefits were validated by two focus groups. Both focus groups confirmed the dashboard’s usefulness, highlighting its ability to reveal optimization opportunities through understandable visualizations. For future work, we aim to support variants and investigate its applicability to other domains.