Bursary Reports 2019
Post-doctoral researcher, Ghent University (Belgium)/ KB National Library of the Netherlands
Workshop: Linked Data for Digital Humanities
I am immensely grateful for participating in the 2019 version of the Digital Humanities Summer School. I learned a great deal on Digital Humanities over the past year and I am still in awe at the wide range of digital applications that are developed to be applied in the Humanities. These make research more accessible, visible and understandable. As a post-doctoral researcher (ECR) in early modern legal/ political-institutional history, I was ‘raised’ with traditional methodology. I am currently working on a project at Ghent University, which (manually) categorises the early modern legislation from the province of Flanders and Holland. At the KB National Library of the Netherlands, I am working on a way to apply machine-learning to this categorisation of legislation. Since I use the same categorisation as the Max-Planck-Institute für europäische Rechtsgeschichte (Legal History) it makes sense to look for means to ensure the use of same definitions within our data sets.
Thus, I attended the workshop Linked Data for Digital Humanities (LD4DH), run by Dr Terhi Nurmikko-Fuller. She is an amazing teacher, with lots of humour (not just in wearing sparkling clothes while discussing SPARQL) and a tremendous amount of experience. She provided us with a step-by-step theoretical background and a lot of chances to get dirty with hands-on experiences. That was not always easy, as Open Source programs make you have a fit – but together with the other teachers, she takes the time to solve our malfunctioning computers.
On Monday, we kicked off with learning the difference between a relational database and Linked (Open) Data, triples,RDF and ontologies. This was somewhat challenging as it feels like learning a whole new language (while English is not my native tongue anyway). We finished our first day with discussing Linked Data in numismatics. It was really useful to have such a practical talk at the end of the day, as it gave some insights into the applicability. Tuesday morning was thought-provoking as we had to write and discuss an ontology for a dataset we were provided with. It was very challenging to think about the data on a meta-level and distinguish between Classes and properties, we had to put this into a tool called Protege (and clean out the given data through OpenRefine). Later on, we learned how this could be applied within Musicology and that entire music pieces can be written using Linked Data. Wednesday had its challenges as many of us struggled to run WebKarma. The program is really useful, as soon as you have your computer run it. When the data is ready, you need to export it to Blazegraph – another interesting program in which you can query your data. I really admire the patience of Terhi, John, and Graham in troubleshooting with many of us, but in the end, managing to solve about every single problem we encountered and – while at it – teach us the use of Web Karma and Blazegraph too! As soon as you have the hang of it, a world of opportunities opens up. An illustration of such an opportunity was given that day by Stephen Downie, with the HathiTrust. Wearing sparkling trousers (on the hottest day ever), Terhi gave us an introduction to SPARQL querying – through which you can query to find whatever data has been linked within your entire data set. Of course, you can also use other programming languages, but for those who do not know, after having worked with all the other tools SPARQL is very functional to start with. Having ‘survived’ all that, we were presented with many other applications – such as Recogito (to find geographical references and plot these on a map) and Knowledge Graph (linking images from the British Museum). The final lecture dealt with the model of CIDOC-CRM – which helped to understand the logic behind data structuring. This nicely wrapped up the challenge that Linked Data poses.
This workshop is extremely useful and practical for those working with datasets that link to other institutes or datasets that should be versatile in its usage (a lot of data of which you know you will need in various combinations). If you want your computer to just do as you order it, you are in the wrong place because the Open Source tools may need some special attention and care before they work the way you expect.
It was a great week to meet like-minded people from various disciplines, countries, and stages of research. Keble College is a fantastic stage (though our room was ‘somewhat hot’), marvellous lunches, and a wide arrange of optional presentations/ discussion fora to attend. A brilliant place to crash-course your knowledge of Digital Humanities in general, and many topics in specific. There is no doubt in my mind that if you want to learn more about Digital Humanities: this is the place to go.