HomeReflection Statement

Reflection Statement

The Linked Open Data course offered by Ashleigh Faith and Aaron Brenner out of the University of Pittsburgh was a well worthwhile adventure into all the aspects of Linked Open Data (LOD) and how it is being employed in the wild. The class offered a breadth of readings, including LOD in the library and academic research space, LOD’s value in proprietary services, where the competitive market dictates success or failure, the use of LOD in automated processing and machine learning techniques, as well as LOD in the wider context of semantic networks. Even though there are clear advantages to using LOD to further knowledge discovery in online collections, the course also stressed that the technology is not a solution for everything. In fact, there are some library and information practices that do not square well with the philosophy behind linked open data, such as the Anyone can say Anything about Any topic (AAA) slogan and the Open World Assumption. Nevertheless, the course offered a grounding in how LOD is important both in curated collections and information retrieval applications. What I most appreciated about this class was the ways in which different topics were explored with well selected readings and an openness for further thought.

To detail the highlights and the challenges of this course, let me present a list of five things I’ve gotten out of the class as well as five points I’d like to explore further.

Five items I enjoyed, learned, explored, and better understood:

  1. The wide application of LOD. I did not realize that standards such as the SSN Ontology existed or that there was already such interest in using sensory data in LOD applications. I can image linked data being the future of applications and services that rely on sensory data to make sense of our world and our personalized experience in it, whether it be for medical or recreational reasons. I liked how the course was quite exploratory in this sense: offering windows into the newest applications of LOD.
  2. The challenges the library community faces in adapting its metadata standards to LOD understood in the wider sense. A week in the course focused on the challenges that the library community faces with its own reliance on RDA and the ways it creates friction with what’s emerging in LOD in other non-library information resources. This was enlightening.
  3. Resources to explore LOD with practice. The course offered a great video on how to explore Wikidata by using example queries to build your own. This is a great way to introduce the topic to people and allow them to explore. The video also showed some really cool applications of LOD, such as Histropedia.
  4. LOD in the real world. Week 6 exposed the students to concrete examples of LOD in the real world. We were encouraged to explore really cool projects underway at the Linked Data for Libraries (LD4D) Gateway. This is a great resource to come back to and learn advancements in the field.
  5. An exposure to ontology work and the ways in which it leverages linked data. The final project I chose focused on using LOD to enhance an ontology for exploring relations inherent in a collection. The process of hands-on practice with ontology work and its use of LOD was a great experience that summed up a take-away lesson of the course, namely that LOD is a public activity that can enhance discoverability not only within one’s own collection, but also within the wider context of the Web and its information resources. I think studying OWL2 would be a great way to advance my knowledge in this subject and continue to work with Protégé.

Five items to learn more about:

  1. Ontologies. I found the project only scratched the surface of the ways in which ontologies assist knowledge discoverability. I plan on becoming more familiar with the OWL 2 standard published by the W3C and using Protégé to see how such standards can embed real word inferences into models.
  2. SPARQL. Though the course offered great readings and exercises on how to query LOD, I feel there could have been more practical experience. Learning a language is all about practice, and I plan on spending some time with Wikidata to understand how they have modeled the world and with the book Learning SPARQL by Bob DuCharme.
  3. Machine Learning. I am intrigued about the use of ML in the work of LOD and ontologies. There were some great readings introducing the use of ML in expanding ontologies to better suit user bases, but this seems like a skill set that’s becoming more and more essential to information retrieval.
  4. Publishing LOD. OpenRefine is a tool that can publish LOD, and it was discussed in one of the class’s required readings. I would really like to explore this open source tool more to see how it can prepare data for cleaning and publication to the Web.
  5. Technology background. We learned how to explore a number of SPARQL endpoints in the class, and the breath of exposure was great. I feel there is much more to learn about the ways in which LOD on the Web is physically implemented. There are a bunch of technologies that involved the client-server relation that were sort of taken for granted in this course. I think exploring the ways in which different types of data is represented on the Web, where it’s stored, and how it’s accessed in a physical way would do a lot of good in terms of understanding the underlying technologies of LOD.

Upon final reflection, I feel the course was a beneficial experience into the different areas of LOD and the technologies that support/enhance it. I think the class could have benefited from more directed peer-to-peer and teacher-to-peer collaboration and that videoconferencing would have been key to making this possible. Nevertheless, the course gave a most enjoyable introduction into the many ways LOD can be thought about, and I have a solid grounding in the major topics for further exploration, research, and practice.