Solve for X(ML): Transforming metadata to transform data access

At the University of Alberta, a subscription to data resources from the Linguistic Data Consortium (LDC) presented a cataloguing problem for library staff and access difficulties for users: resources were published in digital and physical formats, but the digital access was difficult to catalogue, and some datasets were available only in physical format. As a result, there was no single place to discover what datasets the library had purchased, and accessing the data involved a different retrieval process for each format.

This presentation describes a project undertaken by staff from three library units to develop a simplified access model for LDC resources, set to be launched in 2019. LDC metadata available from an actively updated OAI-PMH host provided the basis for a transformation workflow to generate MARC records with custom XSLT stylesheets. This allowed the team to think about user goals rather than cataloguing workflows and imagine an access model where users would not need to know anything about data format. In the new access model, each MARC record includes a link to a retrieval request form. Library staff pick the best format for retrieval and deliver the data electronically to authorized users.