Enhancing cultural heritage collections through linked open data: the production of the mythLOD dataset

The formal representation of cultural metadata has always been a challenge, considering the heterogeneity of cultural objects and especially when dealing with experts' interpretation over them. This paper presents an overview of the mythLOD dataset production as the Mythologiae digital collection revalorisation into linked open data format. The research aims then to explore Mythologiae data leveraging semantic web potentialities, focusing over the formal representation of experts' analysis when associating visual artworks (and their interpretations) to literary sources. The workflow of the project consisted of data model definition, data cleaning and entity linking, conversion (from tabular data to graph) and testing activity (domain experts review over competency questions and two designed on-purpose data visualizations). The result is the mythLOD platform, which present the dataset and the detailed documentation of the research. Additionally, the platform hosts the two data visualization spaces which have been implemented - the online catalogue and the storytelling experiment over Aeneid - as a user-friendly testing unit for the dataset and an additional tool for documenting the project and exploring the collection.