Linked data. It’s one of the hottest topics in the library community. But what is it really? What does it look like? How will it help? In this series I will seek to demystify the concept and present practical examples and use-cases. Some of the topics I will touch on are:
- The basics
- Tools for implementing linked data
- Interviews with linked data practitioners
- What can you do to prepare?
In this part one of the series I will give a brief explanation of linked data; then I will attempt to capture your interest by highlighting how linked data can enhance a variety of library services, including cataloging, digital libraries, scholarly data, and reference.
What is Linked Data?
I’m not going to go into the technical detail of linked data, as that isn’t the purpose of this post. If you’re interested in specifics, please, please contact me.
At its core, linked data is an idea. It’s a way of explicitly linking “things” together, particularly on the web. As Tim Berners-Lee put it:
The Semantic Web isn’t just about putting data on the web. It is about making links, so that a person or machine can explore the web of data. With linked data, when you have some of it, you can find other, related, data.
Resource Description Framework is a framework for realizing linked data. It does so by employing triples, which are fundamentally simple (though RDF can become insanely complex), and by uniquely identifying “things” via URIs/URLs when possible. Here is a quick example:
Behind each of those three “things” is a URL. Graph-wise this comes out to be:
courtesy of W3C’s RDF Validator
This is the basic principle behind linked data. In practice there are a variety of machine-readable languages that are able to employ the RDF model, among them are XML, JSON-LD, TTL, and N-Triples. I won’t go into any specifics, but I encourage you to explore these if you are technologically curious.
What will it be able to do for you?
So, the whole idea of linked data is fine and dandy. But what can it do for you? Why even bother with it? I am now going to toss around some ways linked data will be able to enhance library services. Linked data isn’t at full capacity yet, but it is rapidly becoming flesh and bone. The more the library community “buys into” linked data and prepares for it, the quicker and more powerful linked data will become. Anywho, here we go.
I should clarify that all of these examples conform to the concept of linked open data. There is such a thing as linked “closed” (private) data.
Right now the traditional cataloging world is full of metadata with textual values (strings) and closed, siloed databases. With linked data it can become a world full of uniquely-identified resources (things) and openly available data.
With linked data catalogers will be able to link to linked data vocabularies (there are already a plethora of linked data vocabularies out there, including the Library of Congress authorities and the Getty vocabularies). For users this will add clarification to personal names and subject headings. For catalogers this will eliminate the need for locally updating authorities when a name/label changes. It will also help alleviate the redundant duplication of data.
The “things instead of strings” concept noted above rings true for non-MARC metadata for digital libraries. Digital library staff will be able to link to semantic vocabularies.
Another interesting prospect is that institutions will be able to link metadata to other institutions’ metadata. Why would you do this? Maybe another institution has a digital resource that is closely related with one of yours. Linked data allows this to be done without having to upload another institution’s metadata into a local database; it also allows for metadata provenance to be kept intact (linked data explicitly points back to the resource being described).
Linked data will help scholarly data practitioners more easily keep works and data connected to researchers. This can be done by pointing to a researcher’s ORCID ID or VIVO ID as the “creator”. It will also be possible to pull in researcher profile information from linked data services (I believe VIVO is one; I’m not sure about ORCID).
Two words: semantic LibGuides. With linked data, reference librarians would be able to pull in data from other linked data sources such as Wikipedia (actually, DBpedia). This would allow for automatic updates when the source content changes, keeping information up-to-date with little effort on the librarian’s part.
To take this idea to the extreme: what about a consortial LibGuide knowledge base? Institutions could create, share, and reuse LibGuide data that is openly and freely available. The knowledge base would be maintained and developed by the library community, for the public. I recently came across an institution’s LibGuides that are provided via a vendor. To gain access to the LibGuides you had to log in because of vendor restrictions. How lame is that?
Maybe I’m be a little too capricious, but given time, I believe these are all possible. I look forward to continuing this journey in future posts. If you have any questions, ideas, or corrections, feel free to leave them in a comment or contact me directly. Until next time!