Assessing the capability of open government data for process mining

a study on a Brazilian open data portal

Abstract

The growth in Open Government Data (OGD) availability over the last decade offers significant potential for promoting transparency and improving operational efficiency in government. OGD can support the identification of inefficiencies and the extraction of process models through Process Mining (PM) applied to public service tasks. Various OGD portals, maintained by federal and subnational governments, provide access to diverse public datasets that can be explored for these purposes. Applying PM to OGD, however, requires addressing key aspects such as the tabular nature of data, variability in quality and standardization, and the frequent lack of process context. This study proposes a method to classify and evaluate OGD datasets based on their relevance and potential for process discovery. The method was applied to data from the Brazilian Federal OGD portal (dados.gov.br) to investigate whether OGD can be effectively used in PM tasks to identify processes and bottlenecks. The research involved steps of data collection, selection, and evaluation. Datasets were classified based on their potential utility and suitability for transformation into event logs. The results showed that 24% of the sampled datasets were considered relevant for PM, with 23% in dated tabular format and 1% already in event log structure. Moreover, 52% of the datasets addressed public policies, suggesting the potential of PM to reveal inefficiencies in processes that affect citizens. These findings demonstrate that, with structured evaluation criteria, PM can be effectively applied to public data. This research contributes by presenting an empirical and replicable method to support the evaluation and preparation of OGD for PM appli- cations in digital government. The results reinforce the importance of adopting clear assessment protocols to broaden the use of OGD in understanding and improving public processes.