This project involved a comprehensive analysis of a 20GB training dataset related to an invoice processing workflow. Our primary objective was to meticulously identify structural weaknesses and pinpoint anomalous events within the process. These insights were crucial as they could directly lead to increased operational costs for the cooperation whose data we analyzed. To achieve this, we extensively utilized the pm4py library for process mining, and the findings were effectively visualized using Seaborn, providing clear and actionable insights.
Python
PM4Py
NumPy
Pandas
Seaborn