Abstract : In the last years, cultural heritage institutions have been involved in several initiatives in order to exploit digital means to increase their visibility. Galleries, libraries, archives and museums (GLAM) typically own rich and structured datasets developed over many years and organized by domain, which in principle could be easily connected with the databases of other institutions and then made available online to a larger audience. However, several issues need to be faced, for instance the need for a standard format for data sharing, and the lack of technical skills especially in small museums, so that data manipulation and conversion can hardly be achieved. Relevant standardization efforts such as that of Europeana1 go in this direction and contribute to raise museums’ awareness of the importance of knowledge sharing among cultural heritage institutions. The advantages include driving users to new content, stimulating collaboration in the cultural heritage domain, enabling new scholarship through the availability of new digital content, and more generally increasing the relevance of cultural heritage institutions (Oomen et al., 2012).
In this work, we present the process performed to map the metadata from the Verbo-Visual-Virtual Project (Marchetti et al., 2013) to the Linked Open Data (LOD) cloud and the related data enrichment. Although the work was largely inspired by past efforts by other cultural heritage institutions (Szekely et al., 2013; de Boer et al., 2012; de Boer et al., 2013), we face new challenges, partly related to the small size of the collection, with little-known artists and few information available from other online sources, and partly to the integration of Natural Language Processing (NLP) techniques to enrich the metadata. We show that linking the metadata to DBpedia contributes to improving the quality and the richness of the data owned by the museum, but also that small collections with little-known artists present specific issues, e.g., the limited coverage of external resources, that need to be addressed semi-automatically. We make available both the developed ontology and the RDF data set containing the enriched metadata of our collection.