RISIS Linked Data Course : Summary Report
This is a pictorial summary of our 2-day training course on ‘Linked data for science and innovation studies‘ which took place at the Vrije Universiteit Amsterdam on 23-24 March 2017. The course, organized as a mixture of lectures, hands-on and discussion sessions aimed to help researchers understand basic Linked (Open) Data principles and technologies as well as how it can be used in science and innovation studies.
The course started by introducing the participants in terms of their field of study/research, their technical skills and the potential research questions they expected to be able to answer with the Linked Data. The participants were then clustered into relevant groups to team up for the hands-on sessions. The course consisted of 24 participants from all over europe (Netherlands, Spain, France, UK, Belgium, Portugal, Sweden, Germany, Austria, …).
The morning sessions were about Linked Data Principles to understand what Linked Data is , why we need it, what it looks like and how it works. by the end of these sessions participants could understand the difference between URIs and URLs and could read Turtle files. In the hands-on part people were asked to build their own Linked Data, explore it, and link it to the Web of Data. There were also a collaborative game arranged at the end of morning session to review the Linked Data principles and technologies learnt.
The afternoon sessions started with an introduction on Linked Data & Research within RISIS project given by Peter van den Besselaar where he presented some of the features and use cases of SMS (Semantically Mapping Science) platform. Afterwards Rinke Hoekstra gave a presentation on the Linked Data Lifecycle to elaborate on different phases in the lifecycle of Linked Data as well as important considerations that need to be taken into account when dealing with Linked Data. The Research Data Conversion Hands-on helped participants to convert some of their tabular data to RDF using the OpenRefine tool.
The second day started by a discussion on the results of the data conversion done in the 1st day, and to reflect on how people converted their data to Linked Data, what the common pitfalls were, and what the best practices to deal with those issue are. There was also a reflection on the preparation for the 2nd day where participants were asked to watch some material about SPARQL query language.
The morning sessions started by an introduction on SPARQL to learn how we can query Linked Data, what the syntax of SPARQL is and how we can use it on the Web. During the hands-on session, participants tried to write their own SPARQL queries and also use existing SPARQL endpoints (e.g. DBpedia) and tools (e.g. YASGUI) to answer questions. After the hands-on session, we moved to the another room at W&N building for lunch.
The afternoon sessions started by a presentation from Al Idrissou from KRR (Knowledge Representation & Reasoning) group addressing the importance of context when interlinking datasets. He presented the idea behind scientific lenses when linking RDF data for answering a context-specific research question. After that, Rinke gave a presentation on Linked Data Publishing addressing what triple stores are, and how they work. The goal was for people to understand what solutions for storing Linked Data are available, how participants can use a triple store to publish Linked Data on the web; and how to use mapping files to expose relational data on the web. During the Data Publication Hands-on, participants revised their research question(s), made final adjustments to their Linked Data, and tried to publish the Linked Data on the web by installing a triple store.
In the final session, participants were asked to reflect on what they had learnt so far in the course and what they expect to learn or practice more in the final session. Based on the participant’s feedback, the final session was adapted to address the needs of participants.