I’ve been thinking A LOT about linked data over the past few months. To be less specific, I’ve been thinking about ways that librarians can better share library resources and metadata on the open web, and make our resources more discoverable.
We know that Google, to use one search engine example, ranks things on the web more highly when they are linked TO often, and linked to from other highly ranked domains. We also know that Google, and other major search engines, have collaborated to implement and use schema.org markup in their search algorithms.
Last summer, OCLC announced that they were adding linked data to WorldCat, which is already one of the largest sources of bibliographic data on the web. I’ve spent some of my free time over the last six months poking around at their linked data, reading through the Library extension they’ve created to the schema.org vocabulary, and mulling over their approach.
It occurred to me that one baby step toward including linked data in our library sources is to use simple markup. This is simpler than instituting a triple store and converting our records into RDF-compliant formats, and still enables us to achieve one of our (or at least my) major goals: Making our information more discoverable on the web.
I don’t know that I’ve necessarily done this the right way, but I had a really fun time working on it, and I think it’s a great first step. I took a sample page from the application/database I’ve been working on for the last two years, PAPR. I added some bibliographic markup to this page, including URIs that point to VIAF and the WorldCat linked data. I checked out my mark up in a very cool site I just discovered, Live Microdata. And I fiddled around until it seemed right to me.
This is the marked up code that I settled on. There may be some details that I didn’t entirely get right.
<!DOCTYPE html> <div prefix="schema: http://schema.org/ library: http://purl.org/library rdf: http://www.w3.org/1999/02/22-rdf-syntax-ns#"> <div itemscope itemtype="http://schema.org/CreativeWork/Periodical" rdf:type="http://purl.org/library/Periodical" rdf:type="http://purl.org/vocab/frbr/core#Item"> <div id="title-col1"> <h1 class="title-detail" itemprop="name">Anthropology and humanism.</h1> </div> <div id="title-col2"> <h2>Program: <span class="program_title">Western Regional Storage Trust</span></h2> </div> <div id="content-col1"> <ul id="no_decoration"> <li itemscope itemprop="publisher" itemtype="http://schema.org/Organization" itemid="http://viaf.org/viaf/151905971"><strong>Publisher:</strong><a href="http://viaf.org/viaf/151905971"><span itemprop="name">Published by the Society for Humanistic Anthropology, a unit of the American Anthropological Association</span></a></li> <!-- Query for finding VIAF ID: http://viaf.org/viaf/search?query=local.title+%3D+%22Society%20for%20Humanistic%20Anthropology%22+&maximumRecords=100&sortKeys=title&httpAccept=application/rss%2bxml --> <li><strong>Print ISSN:</strong> 1559-9167</li> <li><strong>Electronic ISSN:</strong> 1548-1409</li> <li itemscope itemprop="http://purl.org/library/oclcnum" itemid="http://www.worldcat.org/oclc/28598923"><strong>OCLC Number:</strong> <a href="http://www.worldcat.org/oclc/28598923"><span itemprop="oclcnum">28598923</span></a></li> <!--This link can be auto-generated using the OCLC numbers; the URL structure is consistent--> </ul> </div> <div id="content-col2"> <ul id="no_decoration"> <li><strong>Validation:</strong> </li> <li><strong>Retention:</strong> Dec. 31, 2035</li> <li itemscope itemprop="http://purl.org/vocab/frbr/core#owner" itemtype="http://schema.org/Organization" itemid="http://worldcat.org/registry/Institutions/125641"><strong>Institution:</strong> <span itemprop="name">Stanford University</span></li> <!--Query for finding Registry ID: http://worldcat.org/webservices/registry/search/Institutions?version=1.1&operation=searchRetrieve&recordSchema=info%3Arfa%2FrfaRegistry%2FschemaInfos%2FadminData&maximumRecords=10&startRecord=1&resultSetTTL=300&recordPacking=xml&query=local.oclcSymbol+%3D+%22STFSP%22 --> <li><strong>Facility:</strong> </li> <li><strong>OCLC Symbol:</strong> STFSP</li> </ul> </div> </div></div>
To explain in actual language, I’m declaring that this information is about an item of schema.org type CreativeWork (and also a FRBR item, and a Periodical in the Library extension language). The item has four properties: a name (title), a publisher, an OCLC number, and an owner. The publisher also has a schema.org type of Organization, as well as an identifier (a VIAF URI), and it has a name property as well. The OCLC Number has an id (the WorldCat URI). And I used the FRBR Owner entity to indicate that a particular schema.org Organization owns this item, and the owner has an ID (the WorldCat Registry URI).
I’m sure this is imperfect, but for a first attempt, it makes me happy. You can check out the JSON representation of this markup at the Live Microdata site.
I would LOVE to hear any thoughts you might have about this approach. Did I get something wrong? Do you have ideas for enhancing and bettering it?
Comments
4 responses to “Linked Data Baby Steps”
Love it! Kudos to you for learning by getting your hands dirty.
Thanks! This is totally the kind of project I love working on.
Love it! Kudos to you for learning by getting your hands dirty.
Thanks! This is totally the kind of project I love working on.