Dados abertos governamentais: adaptação e aplicação de uma metodologia para publicar dados enriquecidos semanticamente no estado de Minas Gerais

• Ano de defesa: 2022
• Autor(a) principal: Marcela Pires Estevanovic
• Tipo de documento: Dissertação
• Tipo de acesso: Acesso aberto
• Idioma: por
• Instituição de defesa: Universidade Federal de Minas Gerais; Brasil; ECI - ESCOLA DE CIENCIA DA INFORMAÇÃO; Programa de Pós-Graduação em Gestão e Organização do Conhecimento; UFMG
• Palavras-chave em Português: Dados abertos governamentais; Ontologia; web semântica; Ciência da informação; Ontologias (Recuperação da informação); Web semântica; Representação do conhecimento (Sistemas especialistas); Informações eletrônicas governamentais
• Link de acesso: http://hdl.handle.net/1843/55003; https://orcid.org/0000-0002-7089-3946

With the advance of new technologies on the Web and the increase in the amount of data and information produced, there is a need to apply techniques that assist in the preparation and organization of data for publication. In the scope of Public Administration, for example, data sets can be published on the Web with the purpose of analyzing the epidemiological profile of a population, or to inform the number of homicides per region, among others. It is important that the data is accessible and properly annotated with metadata to broaden its reuse by different types of people and organizations. However, it is common to find data sets published on the Web in heterogeneous formats or without documentation describing them. Situations of this type hinder the consumption of information by different applications, meeting the need for transparency in public administration, improvement of administrative processes, or even, in helping the government to make decisions based on evidence. In this context, it is important to apply techniques that allow the annotation of data to make the knowledge represented by them explicit. Nevertheless, it is necessary to think about how to enrich them semantically both for human consumption and for processing by computers. Thus, when publishing a dataset, it is also necessary to associate it with the conceptual model that defines it. The use of ontologies allows knowledge to be represented in software artifacts that organize classes, properties, objects, and constraints of a specific knowledge domain. This discussion is within the scope of Open Government Data, where it is necessary to prepare the sets in formats that allow automated processing. Consequently, it opens the way to process and extract knowledge to analyze a large volume of data. In this paper, a methodological approach to publish open government data is proposed. For this, a case study was conducted to annotate data from impositive parliamentary amendments in Minas Gerais, using it as a proof of concept. To build the ontology, focus groups were conducted with Minas Gerais servers to validate the knowledge representation and fill in metadata templates to elaborate the artifacts. Then, data is ingested to generate a knowledge graph that is accessed via a repository by the Business Intelligence program. Finally, the semantically annotated data is presented and a way to "navigate" the concepts through competency questions is explored. Therefore, this work contributes a way to continuously annotate and enrich data and publish it on the Web. This increases the reliability and scalability of models that benefit from Semantic Web Technologies.