LINKED: An Intro to the Semantic Web for Medical Librarians and Biomedical Information Professionals.
This is one of two continuing education courses available on Tuesday, October 9. Jonathan Koffel and Layne Johnson from University of Minnesota taught the class. Jonathan and Layne tag-teamed the presentation by splitting up the ideas into digestible pieces. I’ll do my best to synthesize some essential points.
Without going into too much detail, Jonathan and Layne explained that “objects” (for lack of a better descriptor) on the web are linked by an RDF triple, which consists of the a subject, predicate and object (s->p->o). The subjects and objects in the RDF triple are defined with an uniform resource identifier (URI) which look very similar to URLs. These URIs are readable by both machines and humans and it will display the relevant one when accessed. The predicate in the triple connect the subject and object and are drawn from established ontologies. An ontology defines common and controlled vocabulary for a certain group of users to share information. For example, there is an ontology that describes books, one that many describes relationships between people, etc.
Triples can be searched and retrieved commonly by SPARQL, which is sort of like SQL. This was just a small part of the presentation, but I thought this theory is some of the hardest parts to get one’s head around. Linked data is published and collected in repositories and search engine. You can see some uses of linked data at the New York Times, LinkedCt (Linked Clinical Trials) and DrugBank.
In libraries, we’re still working on learning how to use and integrate linked data. Some of the points that Jonathan and Layne presented were that linked can “increase the visibility of library data, allow more creative reuse of library data and integration of outside data into library records” and “improve cataloging efficiency and innovation.” Personally, I saw the power of linked data in research, particularly with something like LinkedCT, where there would be connections made between conditions, treatments and symptoms. It seemed like there would be lots of potential for deep searching for librarians and their patrons. According to Jonathan, there’s an experimental prototype called Semantic MEDLINE that uses linked data. It looks awesome.
Jonathan and Layne showed us the video from OCLC at the beginning of class and it was great overview:Linked Data for Libraries by OCLC.
Overall, I thought it was a great class and very informative. It’s a big topic for such a short time but I do feel like I have grasp on how linked data works and it’s potential. Pretty good for a four hour class.
It was a great first conference! It was wonderful to meet new people and get involved.
A quick thankful note to the catering company and the hospitality committee. I appreciated the catering company’s consideration of people’s dietary restrictions. I generally eat gluten-free (though definitely broke down over the Minnesota beer tasting) and it’s great that there is enough variety for a full meal, whether you eat vegetarian or vegan or gluten-free.