Linked Data & Digital Libraries

Annotations by Kaley DeLong

Agenjo, X., Hernández, F., & Viedma, A. (2012). Data aggregation and dissemination of authority records through linked open data in a European context. Cataloging & Classification Quarterly, 50(8), 803-829. http://dx.doi.org/10.1080/01639374.2012.711441

In their discussion of the Polymath Virtual Library, Agenjo, Hernández, and Viedma explore the process and challenges of using authority records to develop linked data in a multi-language virtual library. They argue that bibliographical design is more important than the technical aspects of instituting Linked Open Data (LOD) (p. 804) and one of the challenges of dealing with bibliographical design is the lack of information about the language in the MARC record. One issue is a lack of accents in MARC records that can change the name of an author based on the language being used. This adds to the complexity of the initial records and the challenge professionals face when developing LOD resources. Aimed at professionals, the authors examine the difficulty of linking authority records to existing LOD resources and working with Eurpoeana. Agenjo, et.al.’s study is helpful for understanding the unique challenge foreign languages play in migrating records into LOD.

Fox, R. (2016). From strings to things. Digital Library Perspectives, 32(1), 2-6. http://dx.doi.org/10.1108/DLP-10-2015-0020

In his article on the developments of the semantic web, Fox discusses several large-scale projects that have attempted to deal with issues of language to better support Linked Open Data. The article is aimed at those familiar with LOD and libraries; however, his brief discussion of the practice of the Internet versus the idea of how it should work may appeal to those outside the practice of digital libraries. Fox’s article does not directly discuss the process of dealing with LOD in digital libraries; instead it discusses recent projects such as MetaWeb’s large scale “knowledge base” (p. 4) and how such projects have attempted to utilize web resources like Wikipedia to build their own database. Fox argues that, in order to have interoperability between identification systems and their data, we must begin with a common description such as OCLC’s schema.org (p. 5). The article is useful for understanding current efforts being undertaken by larger entities to develop Linked Open Data and the importance of finding a common tool that can be used both inside and outside libraries to develop

Hallo, M., Luián-Mora, S., Maté, A. & Trujillo, J. (2016). Current state of linked data in digital libraries. Journal of Information Science, 42(2), 117-127. http://dx.doi.org/10.1177/0165551515594729

The research conducted by Hall, Luián-Mora, Maté, and Trujillo examines the current state of the development of Linked Data applications in digital libraries and the various benefits and problems that are faced by various institutions. Their study looked at five digital libraries; the National Library of France, Europeana, the Library of Congress USA, the British Library, and the National Library of Spain and found that the success of linked data “depends on the ability to identify, reuse, or connect data and data models” (p. 121). They also found that while Linked Data applications helped improve the visibility and recovery of data, the number of vocabularies for the same metadata and a lack of resources (in the form of both tools and experts) impeded progress. The study is helpful for understanding the current issues and benefits of linked data in these five well known digital libraries, as well as where future research is needed to improve both an understanding of Linked Data and its use in digital libraries.

Konstantinou, N., Spanos, D-E., Houssos, N., & Mitrou, N. (2013). Exposing scholarly information as linked open data: RDFizing DSpace contents. The Electronic Library, 32(6), 834-851. http://dx.doi.org/10.1108/EL-12-2012-0156

Konstantinou, Spanos, Houssos, & Mitrou’s article discusses the potential of converting existing digital library data into Linked Open Data through the use of a relational-to-RDF mapping engine, applied to a digital library’s current relational-database backed systems (p. 835). To test this application, the authors present their case using the DSpace platform. Using this application, the authors were able to populate a database with DSpace entries while still allowing the triplestore to coexist with the current version (p. 844). The article provides a guide for mapping non-legacy Dublin Core elements that act as a guideline for other mapping activities. Their process provides an approach that allows a digital library to avoid vendor issues by being widely applicable, as well as asynchronous and non-obtrusive. The article is helpful for understanding a potential solution to the wide variety, and potential lack of cohesion between, languages and technology that digital libraries face when dealing with Linked Data. It provides a technical approach for digital librarians looking for solutions to supported Linked Data without the erasing previous efforts and without investing in new vendors.

Mathews, E. & Smart, L. (2016). Piloting linked open data on artists’ books: A case study in interoperability and sustainability. Digital Humanities and the Visual, 43(2). Retrieved from http://online.vraweb.org/vrab/vol43/iss2/5

Mathews and Smart’s article discusses the process of publishing Linked Open Data for digital surrogates of artists’ books. They provide a thorough discussion of the process they completed, as well as where they struggled with the scaling of their metadata work. The article provides a useful set of questions and guidelines for professionals looking to engage in similar projects. For a discussion on Linked Open Data and digital libraries, this article is helpful for understanding the issues that arise the practice of developing LOD and the unique challenges faced by a specific type of work (in this case, artist books which incorporate various traditions such as books, painting and performance). The source is also one of the few that indicates the difference between Linked Data and Linked Open Data and the issues associated with this distinction in relation to their project.

Mitchell, E. (2013). Three case studies in linked open data. Library Technology Reports, 49(5), 26-43.

This article examines two digital libraries, the Eurpoeana digital library and the Digital Public Library of America, as well as the Library of Congress’s BIBFRAME initiative. Mitchell uses a case study approach to examine the three different entities’ data models, content rules, metadata schema and vocabularies. In addition to discussing the general rules for each entities’ metadata, Mitchell also provides graphical visualizations of specific items to show how each is handled. The article is heavy in its technical discussion and is directed towards those already familiar with Linked Data and coding such as XML, JSON, and APIs. The discussion of each digital library’s handling of metadata is helpful for seeing the range of approaches that can be taken by various institutions, as well as understanding the diversity of technical approaches available for dealing with Linked Data.

Silvello, G. (2014) A methodology for citing linked open data subsets. D-Lib magazine, 21(1/2). Retrieved from http://www.dlib.org/dlib/january15/silvello/01silvello.html

In this study of data citation and Linked Open Data, Silvello focuses on using the process of experimental evaluation of search engines to develop a methodology that can be customized, create citations with variable granularity and create references that are both human and machine-readable. Silvello’s methodology uses meta-graphs that help to show the explicit connections between statements and identifiers by building off of RDF graphs. The layering of these graphs allows for both machine and human readability and alteration, an important factor for the development of Linked Data in digital libraries. This methodology allows for more precise experimentation and evaluation by helping to establish clear links that can be quantitatively measured. Aimed at professionals experienced with RDF and Linked Data, the article is unclear about some basic terminology. However, its frequent presentation of the visuals of the meta-graphs is useful for understanding the steps laid out by Silvello’s methodology.

Solodovnik, I. (2011). Metadata issues in digital libraries: Key concepts and perspectives. Italian Journal of Library and Information Science, 2(2), http://dx.doi.org/10.4403/jlis.it-4663

In this article on metadata issues in digital libraries, Solodovnik discusses Linked Data as part of the larger issue of dealing with metadata within a digital library. Solodovnik prefaces her review of various Linked data applications with a discussion on the current use of RDF and the need for non-library interoperability. She makes note of the recommendations from AIMS (Agricultural Information Management Standards) and their attempt to inform how bibliographic Linked Data is published and consumed. This article is the broadest of this bibliography and would serve as a strong starting point for understanding the basics of various metadata issues and where Linked Data may fulfill a need for metadata that is more ontology-based than previous efforts.

Southwick, S. (2013). Exposing missing links: From CONTENTdm digital collections to the linked open data cloud. Presentation at Best Practices Exchange, Salt Lake City, UT. Retrieved from http://digitalscholarship.unlv.edu/libfacpresentation/110

This presentation by Southwick discusses the practice of converting the metadata of existing digital content collections to Linked Open Data. Southwick’s presentation is aimed at other professional librarians; however, she reviews the basic concepts of linked data as part of the overall presentation. Southwick’s presentation is useful for understanding the basic deconstruction of metadata into triples (subject-predicate-object) that is necessary for the functioning of linked data. The graphical representation of the deconstructed metadata is a helpful tool for understanding the structure of the Semantic Web as it is built from a specific record. The presentation’s discussion of using OpenRefine and Mulgara is useful for those unfamiliar with these applications.

Southwick, S. (2015). A guide for transforming digital collections metadata into linked data using open source technologies.Journal of Library Metadata, 15(1), 1-35. http://dx.doi.org/10.1080/19386389.2015.1007009

This article by Southwick follows up on Southwick (2013)’s presentation on University of Nevada, Las Vegas Library’s project to transform their digital collection’s metadata into Linked Open Data. In addition to covering the basic aspects of Linked Data, Southwick’s article also discusses the UNLV project’s approach, highlighting both the goals and research questions associated with the project. The article discusses the planning phase of the conversion; however, the article focuses more on the process than on the individual experience of the UNLV staff. This makes it a useful article for understanding how a similar project may be conducted in a different environment and what aspects must be considered when doing so. The article’s focus on the practical aspects of the process, over the theoretical, make it a useful tool for understanding the process and specific tools used when creating Linked Data in a digital library.

Yoose, B. & Perkins, J. (2013). The linked open data landscape in libraries and beyond. Journal of Library Metadata, 13, 197-211. http://dx.doi.org/10.1080/19386389.2013.826075

Yoose and Perkin’s article focuses on the basics of Linked Data and Linked Open Data and the major projects with which they are associated. The article is one of the few to lay out the differences between the Semantic Web and Linked Data and this makes it helpful for readers who are unfamiliar with the subject of Linked Data or its application in digital libraries. Like many other articles that deal with Linked Open Data, Yoose and Perkins discuss Tim Berners-Lee’s LOD star ranking system that specifies levels of openness, from web availability (1-star) to Linked Data that uses open standards from W3c, is in a non-proprietary format, is machine-readable, and is available on the web (5-stars). The discussion of the various tools like OpenRefine, the different registries and hubs like DBpedia, and the current library/museum initiatives like the BIBFRAME and Eurpoeana, make this article a useful primer for other articles on Linked Data in digital libraries.