Do DFG à rede de Petri: descoberta, conformidade e análise de desempenho em logs reais de processos

Abstract

This undergraduate final project investigates the application of Process Mining techniques in institutional settings by proposing a reproducible pipeline for extracting, curating, and transforming administrative records into analyzable event logs, with an initial modeling stage based on the Directly-Follows Graph (DFG). The work discusses typical challenges of real-world datasets – such as heterogeneity, incompleteness, and the lack of consistent identifiers – and emphasizes how the proposed procedure can be adapted to other datasets described in the databases chapter.Although the case study using the Central de Emendas da- tabase was the initial goal, time constraints and strict accreditation requirements prevented data access within the project timeframe, leaving this step for future continuation. As next directions, the thesis outlines a complete application under accreditation, the institutionalization of curation/anonymization routines, and the move toward more expressive models such as Petri nets, including perspectives on formal verification with automated reasoners and theorem provers.