Presenter
Chiara Carlucci (@Chiara_Carlucci)
Valeria Cioccolo
Abstract
This contribution presents IN.LOD, a project developed at INAPP to transform the semantic infrastructure of an OPAC into a Linked Open Data environment, making institutional knowledge more interconnected, reusable, and easier to discover. The current implementation focuses on the INAPP Thesaurus, with around 2,500 terms modelled as semantic entities that connect bibliographic resources with concepts, people, organisations, places and external sources. The transformation is supported by an automated pipeline: Thesaurus data are exported from the OPAC as CSV files and converted into RDF/Turtle through dedicated mapping files, then published in a Virtuoso triplestore. The process can be run manually or scheduled, with logging to monitor successful transformations and errors. IN.LOD uses URI-based modelling, controlled vocabularies and the semantic relationships defined in the Thesaurus. Links to Wikidata, ORCID and GeoNames are generated automatically and checked through manual sampling. The resulting knowledge graph contains more than 12,000 semantic entities and over 56,000 relationships. The project also addresses the sustainability of the semantic infrastructure, building on resources such as schema.gov.it and Wikidata and considering identifiers, vocabularies, and the maintenance of external links over time. A formal usability study has not yet been carried out, but it is planned for a later phase. Beyond the catalogue, IN.LOD raises a broader issue concerning semantic infrastructures and generative AI, since well-structured, identifiable and interconnected information can support more reliable machine discovery and retrieval. This also underlines the evolving role of librarians and knowledge managers, whose work increasingly includes modelling entities and relationships, maintaining vocabularies and links, and providing context for both human users and AI systems.