ProRCA: A Causal Python Package for Actionable Root Cause Analysis in Real-world Business Scenarios

Modern operational landscapes, spanning domains such as retail, healthcare, finance, and software systems, are increasingly characterized by complex interdependencies and massive data streams. In such settings, anomalies rarely arise from a single isolated factor; rather, they emerge as the cumulative effect of multi-hop causal chains. Existing RCA methods typically focus on detecting outliers or isolating single nodes based on correlation or localized attribution. However, these approaches do not provide a complete explanation of why a failure occurred. In other words they do not systematically trace all possible causal pathways from an observed effect back to its initial triggers. The primary motivation for our work is to address this limitation by developing a package that systematically reconstructs the full causal pathway from an observed anomaly back to its root cause. By leveraging the strengths of the DoWhy causal inference library, our method extends existing techniques to not only identify individual anomalous nodes but also trace entire multi-hop causal chains. This end-to-end approach enables practitioners to intervene precisely at the earliest disruption points, thereby reducing the risk of recurring failures and improving overall system reliability.