Entidades Em Documentos De Arquivo E Sua Expansão Com Fontes De Dados Ligados No Projeto EPISA

• Main Author: Camilla Oliveira da Silveira
• Publication Date: 2022
• Format: Master thesis
• Language: por
• Source: Repositórios Científicos de Acesso Aberto de Portugal (RCAAP)
• Download full: https://hdl.handle.net/10216/143034
• topic: Ciências da comunicação; Media and communications

The Portuguese National Archives (Arquivo Nacional da Torre do Tombo) is an example of an institution that maintains a collection of great relevance to the Portuguese national memory. It holds original documents from the 9th century to the present day, many already digitized and available to be searched by researchers and the general public. The archive has been adapting to embrace an open perspective in which archival data is linked to global information sources, following a worldwide trend of making information from cultural heritage repositories more accessible. This work analyzed a sample of 25 records from the Portuguese National Archives, chosen by archival specialists as representative of different fonds and description levels, to identify entities and properties and explore relationships with other non-archival resources. The goal is to provide additional information to the archival interfaces. After selecting the entities in the records, Wikidata, DBpedia and Europeana databases were chosen for data enrichment and linking. The analysis of the entities in the records provided hundreds of properties directly related to classes in the corresponding databases. A more significant set of properties was obtained by selecting properties that had more data entered into the databases for entities of the corresponding classes. These entities and properties were then subjected to tests using Wikidata Query Service, DBpedia SPARQL Explorer and the Europeana Virtuoso SPARQL Query Editor. The queries performed with the entities in Wikidata Query Service and DBpedia SPARQL Explorer only provided results for certain properties, since not all entities have data in the database. On the other hand, the properties tested proved to enable relevant data about the entities, providing additional information about the entities they are connected with. However, with Europeana the entity searches did not prove productive, and it was not possible to retrieve information about entity properties as performed in the previous databases. Europeana is a great aggregator of metadata, however its metadata is still in the process of being converted into linked open data. Finally, further studies on repositories are needed in order to investigate the feasibility of exploring the links between archival resources and information in large open archives. Keywords: archives, linked open data, semantic web