Digital Scriptorium
Give a short description of your project, specifically mentioning why you chose to use Wikibase.
Digital Scriptorium is a member-supported consortium representing American libraries, museums, and other cultural heritage institutions. It maintains an online national catalog of premodern manuscripts so as to enable cross-institutional searching. Wikibase is an open-source, low-barrier, out-of-the-box solution for managing and publishing our contributed member data, making it a one-stop shop for our Linked Data project: a knowledge graph, a user interface, automated data management tools and a SPARQL query interface. Further, the cloud-based service affords us flexibility as our project leadership changes overtime.
Describe how you cultivated a community of contributors for your Wikibase project.
We manage the content on Wikibase that our contributors give to us to transform. Although contributors do not have direct access to make edits in Wikibase, DS member institutions are welcome to provide us with feedback about the accuracy and presentation of their data. The added value we provide in harmonizing, enriching, and publishing member metadata as LOD has not only encouraged current members to contribute but has garnered interest from new and prospective member institutions. The ability to query data using SPARQL is also a benefit to the scholars and researchers who use the database.
Describe the process you went through when deciding how to model your data in Wikibase.
We surveyed data points that members were using to describe manuscripts in their collections in various cataloging formats, identifying a simple schema based on common data points. We then adapted that schema to a Wikibase model by identifying relevant entities and properties, which included qualifier properties to enrich semi-structured metadata with authority values. The full process is described in our research paper here.
Did Wikibase meet your needs? Describe the challenges and successes you faced when implementing Wikibase.
We were able to accomplish our goals in a relatively short period of time because we didn’t have to build this stack ourselves. We don’t work with Wikibase and PHP locally, and Wikibase.Cloud has been invaluable because we don’t have to manage the system ourselves. While Wikibase does provide some LOD capabilities, its model diverges from RDF. Most notably, the property-qualifier arrangement simplifies the presentation and management of linked (small-L) data in the user interface, but it adds complexity to querying and the management of data outside the Wikibase environment. If we choose to migrate our data to an RDF model, it will require significant data transformation and reworking. We are also finding ways to work around performance and reliability issues with the query service, which can update slowly and does not always return all items specified by a query-