Preparing for the post-MARC world:
BIBFRAME Transformation for Enhanced Discovery
Innovation Fund request by: Qiang Jin and Jim Hahn
- The project idea
The end result of this project is enhancing discovery of library data – bringing like sets of content together in more contemporary and easy to understand views assisting users in locating sets of associated bibliographic metadata. Referencing previous work in data mining of bibliographic metadata (Hahn & Diaz 2013) this grant would extend the work to focus on the mining of like bibliographic works and instances when users search library discovery systems at the University of Illinois.
- Objectives: what problem(s) will it solve
Discovery of similar sets of content in the library system is of critical concern. Having large sets of collections in the library catalog has led to increased challenges in searching and selecting the item and sets of related items that users desire. In fact, the current data structures in VuFind do not easily allow the representation of BIBFRAME views of the entity sets (these are sets of associated works, instances, authorities, and annotations).
This is a problem since a BIBFRAME-like representation for library collections could offer users a deeper picture of the bibliographic works and related entities they are searching. In essence, this project looks to address shortcomings of traditional MARC data structures. Further, the project would be contributing to next generation discovery advances by contributing our work to the BIBFRAME Implementation Register, which will enhance the status of the UUC Library at the national level.
- How it fits with existing activities in the Library
Having consulted with others doing library catalog data extraction, we see a need to offer alternative structures to the XML data being pulled from the catalog. With a flat XML view of library data the library does not have the ability to see related sets of editions and manuscripts or derivative works. Therefore, this work fills a gap between where the optimal catalog views should be and where they currently are.
As the CARLI I-Share consortium searches for a new ILS we view this work as complimentary to their investigation since next generation discovery views should offer the contemporary library data models (post-MARC catalogs) we intend to prototype in this grant.
Utilizing the Easy Search XML gateway, we would pair BIBFRAME feeds with the results from federated search and plan in our work a bento module (or component) to add to the library discovery system.
- Resources needed (be specific, e.g., for backfill, equipment, hourly employees)
10 hours a week graduate hourly student for three semesters: Fall 2014, Spring 2015, and Summer 2015. We have identified potential CS graduate students we could recruit for this work.
We can utilize already existing server space allocated for prototyping.
- Sustainability (for Innovation Funding)
The BIBFRAME transforms and bento development portion of this work may find a home in the Discovery department or may also be useful for Metadata and CAM departments. The development BIBFRAME competencies by additional library units would offer sustained innovation using the technologies and processes utilized over the grant period.
Full catalog data transform using MARC to BIBFRAME transformation service and corpus of XML catalog data.
Using transformed data, produce a RESTful web service powered by Blacklight engine that would power a bento module offering BIBRFRAME views of catalog data. Code available here: https://github.com/projectblacklight/blacklight/wiki/JSON-API
Incorporate bento-like module pulling from the Blacklight JSON API into the discovery engine (Easy Search, or other). This would complete the task of adding BIBFRAME encoded views to library data.
- How to measure benefits of the project
Our assessment will include a systematic measuring between the stated goals and the outputs of the grant work. Specifically the benefits can be measured in the associations of bibliographic entities that we are able to identify in data processing, the web services we build using Backlight as the data engine, and the users who are better able to view associations of bibliographic data (related works, instances, authorities, and annotations) from a single discovery tool.
- How to determine whether the project has succeeded or failed
This project will be successful by contributing a module to bento views of catalog data. Lessons learned in BIBFRAME transformation will prepare the library to be leaders in this area as we contribute to the implementation of post-MARC discovery systems.
Hahn, J. & Diaz, C. (2013). Formative evaluation of near-semantic search interfaces. Internet Reference Services Quarterly 18 (3/4).
Bibliographic Framework as a Web of Data: Linked Data Model and Supporting Services http://www.loc.gov/bibframe/pdf/marcld-report-11-21-2012.pdf