A workflow for publishing Linked Data for conservation documentation: A case study from Oxford Museums and Libraries

Linked Data is a set of technologies which allow large scale integration of datasets. With the term 'integration' we refer to the process which results in a unified and distributed dataset that enables querying data from different systems jointly. A cornerstone of such process is an 'ontology': a formalised document which describes the discourse in a domain through basic concepts and relationships of concepts. In 2017 Oxford University ran a pilot project to assess the potential of Linked Data for integrating datasets from Oxford collections. The Oxford datasets included data from conservation studios at the Bodleian Library and the Ashmolean Museum. In this paper we present a workflow for transforming existing conservation data from different sources into Linked Data. Our workflow is system independent and accommodates different levels of data quality. We address issues about the technical publication of conservation terminology for Linked Data applications with particular emphasis in Universal Resource Identifiers (URIs). We discuss frequent characteristics of conservation data and how these are expressed using a heritage ontology: the CIDOC-CRM. We continue with observations around enriching conservation data. We conclude with an assessment of the benefits of Linked Data integration in conservation, including the new knowledge that can be extracted from integrated data, the shift from documenting records of single objects to documenting collections and the evaluation of treatments in conservation in the same fashion as clinical studies. The paper will be illustrated with examples from Oxford datasets and a demo system querying integrated data.