Linked data ontologies use principles of link popularity as a mechanism for algorithms to rank topical search. Increasingly, they form the basis for Discovery layers running on library metadata in contrast to retrieving specific MARC fields in weighted combination with others. There are also concerns about privacy and reductionism when categorizing content in ontologies, especially content related to protected identities.

In distinguishing between linked data ontologies and ontologies embedded in MARC metadata such as LCSH, Emily will demonstrate a methodology using SQL on MARC pulled from WMS query collections to capture “out-of-ontology" content. By combining subject headings from embedded MARC fields, such as LCSH, with other data points, librarians can retrieve materials on a subject that may not be related to others in an explicit or easily defined way, generating collection insights to potentially remediate gaps in representation while respecting the privacy and varied experience of persons with protected identities.

Presenter:

• Emily Fidelman, Head of Metadata, West Virginia University