Shout LOUD on a road trip to FAIRness: experience with integrating open research data at the Bibliotheca Hertziana

Abstract

Modern-day research in digital humanities is an inherently intersectional activity that borrows from, and in turn contributes to, a multitude of domains previously seen as having little bearing on the discipline at hand. Art history, for instance, operates today at the crossroads of social studies, digital libraries, geographical information systems, data modelling, and cognitive computing, yet its problems inform research questions within all of these fields, which veer towards making the output of prior research readily available to humanists in their interaction with digital resources. This is reflected in the way data are represented, stored and published: with various intra- and inter-institutional research endeavours relying upon output that could and should be shared, the notion of ‘leaving the data silo’ with a view on interoperability acquires even greater significance. Scholars and policymakers are supporting this view with guidelines, such as the FAIR principles, and standards, such as Linked Open Data, that implement them, with technologies whose coverage, complexity and lifespans vary. A point is being approached, however, where the technological opportunities permit a continuous interoperability between established and concluded data-intensive projects, and current projects whose underlying datasets evolve. This enables the data production of one institution to be viewed as one harmonically interlinked knowledge graph, which can be queried through a global understanding of the ontological models that dominate the fields involved. This paper is an overview of past and present efforts of mine in the creation of digital humanities knowledge graphs over the past decade, from music history to the societal ramifications of the history of architecture. This contribution highlights the variability of concurrent research environments at the Bibliotheca Hertziana, not only in the state of their activities, but also in the ways they manage their data life-cycles, and exemplifies possible combinations of FAIR data management platforms and integration techniques, suitable for different scenarios resulting from such variability. The paper concludes with an example of how feedback from the art history domain called for novel directions for data science and Semantic Web scholars to follow, by proposing that the Linked Open Data paradigm adopt a notion of usability in the very morphology of published data, thus becoming Linked Open Usable Data.