Cutting-edge research from the University Libraries is creating a new way to discover and link information online.
Cory Lampert and Silvia Southwick work on linked open data research.
Former U.S. Congresswoman Shelley Berkley is the daughter of George Levine, who was employed by Sands Hotel and Casino, where Jack Entratter worked as an executive and show producer.
Expressing the links between people, places, and things is at the core of research on linked data. Digital Collections Head Cory Lampert and recently retired Metadata Librarian Silvia Southwick have led UNLV University Libraries’ research on linked open data for libraries.
“I first became interested in linked open data after attending a presentation on it from Karen Coyle, a digital libraries consultant who is widely considered to be a leader in research on bibliographic data,” said Lampert. “Around the same time, Su Kim Chung moderated a panel on linked open data for the Society of American Archivists. We saw a lot of interest in the topic within the library community and wanted to see how we could use linked open data to promote our collections.”
In April 2012, the University Libraries formed the UNLV Linked Open Data Study group, consisting of representatives from many libraries divisions, including technical services, web development, special collections, library technologies, public services, and assessment. The goal was to gather information and brainstorm ways that UNLV could tap into the emerging field of linked open data.
After gaining a better understanding of linked open data concepts through the study group, Lampert and Southwick were convinced that linked open data concepts held great potential to expose the rich connections between the libraries’ many special collections information resources. Where current digital libraries often fell short in revealing how two items from different collections might be related, linked open data offered a set of principles to express and communicate this context to users. With this research question to guide their hands-on work, they officially launched the UNLV Linked Open Data Project in 2013.
“Our goal was to study the feasibility of developing a process that would enable us to transform our collections into linked open data while preserving all the information contained in the original records,” said Southwick. “Then we wanted to see how we could improve the discoverability of our digital collections and create connections with other related data sets on the Internet using linked open data.”
How linked data works
Linked open data is just one aspect of the larger semantic web. Large institutions, like the Library of Congress, that contain vast quantities of information are working to structure and share their data in a way that is machine-actionable, much in the way that they have led other national data initiatives.
The University Libraries is using open linked data research to connect its unique collections with these large datasets and the larger linked open data cloud.
“Each object, as well as associated statements about the object, is assigned a unique identifying number - like a Social Security Number - that are used to match with the data on the same entity in other information repositories around the world,” said Southwick. “This creates a vast interconnected web, and our information becomes part of that web.”
These numbers are connected using a linked open data “triple” - a subject, predicate, and object - to create a simple sentence. These keywords are embedded into the existing metadata for each of the datasets.
“Every object, person, business, or thing can be related to another thing using these triples,” said Southwick.
This transformation of the data makes it more “machine friendly,” thereby increasing its reusability, discoverability, and accessibility beyond current University Libraries users.
As research on linked open data becomes more sophisticated and standardized, more institutions are able to link to this web, exponentially expanding the resources and information that are available for students, scientists, academics, researchers, and the general public.
“This will be the way that people search for information going forward,” said Lampert. “The process exposes context and the relationships between things instead of providing you with millions of search results that are triggered by a keyword. Instead, smarter results sets can save the user time and provide them new ways to dive into just the aspects of the data that are meaningful to their research.”
Presenting the findings
In the ensuing years since the launch of the research project, Lampert and Southwick have published and presented extensively on linked data.
In October 2013, they presented a full-day training workshop, “Not Just for Geeks: A practical approach to linked data for digital collections managers,” for the Mountain West Digital Library partners, a collaborative of universities, colleges, public libraries, museum archives, and historical societies from six states. In addition to presenting their research, Southwick and Lampert offered hands on exercises demonstrating how to transform digital collections metadata into linked open data.
“By 2014, we were able to deliver presentations that were focused on the tools and technologies that institutions could use to implement linked open data effectively,” said Lampert. “As we continued on in our research, we were able to publish on guiding principles for creating controlled vocabularies and normalizing data so that it could be easily linked with other data sets.”
Lampert and Southwick wrote a book chapter, “On the Brink of Linked Open Data: Evolving Workflows and Staff Expertise,” for “Cutting-Edge Research in Developing the Library of the Future: New Path for Building Future Services,” which was released in 2015. Lampert has also written a book chapter with Digital Special Collections Librarian Emily Lapworth and former Digital Collections Specialist Meghan Gross on “Optimizing Merged Metadata Standards for Online Community History: A Linked Open Data Approach,” which will be published this year.
In addition, Lampert and Southwick have collectively published four journal articles published in the Journal of Library Metadata; given 11 invited presentations, webinars, and training sessions for organizations like the Association for Information Science and Technology and the Art Libraries Association of North America, and five refereed conference presentations for the American Library Association, Coalition for Networked Information, Virtual CONTENTdm User Group, and the Association of College Research Libraries.
Cory Lampert and Silvia Southwick with a linked open data map.
Putting the research to work
Digital Collections has used the linked open data research to develop a new online browsing tool that will make it easy to navigate through the university’s digital collections, while also connecting to libraries, universities, and other information repositories around the world.
Lampert and Southwick have tapped the University Libraries’ Southern Nevada Jewish Heritage Project as a pilot project for “Navigator.” The online browsing tool uses the metadata embedded in the files to create visualized linked between collections.
“When you use online search engines to conduct an internet search, you get a long list of search results without any context,” said Lampert. “With non-proprietary linked open data, there is no ambiguity and connections between people, organizations, and communities are automatically linked for users, which leads to richer discovery of information.”
When the University Libraries transform traditional metadata into linked open data, the new application, Navigator enables users to see connections using the triples.
“This method provides exact results and eliminates the need for users to review millions of results just to see if they contain information related to their search topic. It eliminates the ambiguity,” said Southwick.
The visualization tools of Navigator show the interconnectivity between the many individuals, organizations, archival collections, and digitized objects that make up the Southern Nevada Jewish Heritage Project.
“Archivists and librarians have been aware of the importance of context in the materials they manage,” said Lampert. “But often the relationships are implicit. By pursuing linked open data, the University Libraries is communicating these important connections with users.”
Like the Berkley example, users can see more of what we already know about how people are connected and how their relationships have evolved over time.
A five-year partnership
Though she is leaving the University Libraries behind for time with family and travel in retirement, Southwick hopes to see the UNLV project continue to grow and the university’s collections become linked in the vast online web of information.
“The project opens up our data for further discovery,” said Southwick. “The true realization of this project’s success will come as we continue to connect to other repositories with more of our data published online and made easily discoverable.”
Her contributions to the project will live on through the Southern Nevada Jewish Heritage Project and future Digital Collections endeavors using Navigator. As a co-lead on the project, Lampert will continue to build the linked open data research at UNLV and further develop the Navigator interface.
“We really want to be part of leading and empowering others in the adoption of linked open data in cultural heritage institutions,” said Lampert. “It has been my great pleasure working on this project with Silvia these past five years to develop our linked open data research agenda and launch the Navigator interface for public use. Scholars like Silvia are rare and her expertise will be sorely missed, but we are wishing her well in her much-deserved retirement.”