Linked open data in authoring virtual exhibitions

In the last years, virtual exhibitions have been widely adopted to enhance traditional museums and enable active interaction with culture without posing any physical constraints. Nevertheless, people interested in cultural heritage still behave as visitors. To fully engage them, we propose to let cultural heritage lovers play the role of exhibition curators. In authoring virtual exhibitions, users have to perform a data selection phase that poses several challenges, including finding data sources and extracting data of interest. We aim to take advantage of data published as Knowledge Graphs in the Linked Open Data format. Users can query geographically distributed artworks thanks to their linking nature, manipulate heterogeneous data, and easily customise their exhibitions by exploiting the wide range of available cultural heritage knowledge graphs. However, the complexity of linked open data query languages (such as SPARQL) threatens their exploitation. Consequently, we need to mask SPARQL technical challenges and guide users in naturally posing questions to unlock the potentialities of linked open data to a broader audience. We propose a virtual exhibition authoring tool that guides users from knowledge graphs querying to the automatic generation of virtual experiences. The Knowledge Graph query phase relies on ELODIE, a natural language interface to scaffold users in retrieving data of interest without asking for technical skills in query languages. We introduce our prototype by describing its operating mechanism and by detailing its components. We present a Van Gogh's experience as a use case by collecting all the artist's artworks published on DBpedia (a well-known and general purpose knowledge graph) and organise them in a virtual reality-based virtual exhibition. Finally, we conclude by overviewing advantages and technical challenges posed by linked open data in designing and developing knowledge graph exploitation tools.