Automatic extraction of the process of changes to legal acts

Abstract

In an era of frequent legislative changes and a growing volume of published laws, automating amendment-related processes is essential for effective analysis and interpretation of legal documents. Manually tracking and describing relationships between new and ammending acts in the Journal of Laws of the Republic of Poland is time-consuming and prone to error. These tasks can be supported through metadata extraction tools and process mining algorithms. This work presents a system for automatic extraction and visualization of amendment processes. Parsing tools were applied to obtain and process the texts of laws, transforming the extracted metadata into event logs and a Directly-Follows Graph (DFG) for process mining. A web-based application provides graphical representation of the data. The resulting tool automatically generates a graph illustrating relationships between acts and their evolution over time. A survey of legal professionals and non-expert users confirms improved efficiency and readability in identifying changes and visualizing legislative paths. The tool offers practical support for lawyers and analysts working with historical and current legislative processes.