Wrap-up Creative Camp #3

The third and final Creative Camp was hosted by the National Library of the Netherlands and took place from 26 to 30 October 2015 in The Hague. The goal of this last ‘hackathon’ was to connect various datasets from and about governments, ranging from parliamentary proceedings and policy documents to news media and social media, and work towards a broad, international, cross-media network of ‘Linked Government’ data. The week was concluded with a public symposium on Friday, titled ‘Open Data for the Social Sciences and Humanities’.

The call for participation was spread in June 2015. The selection process for the final camp was strict. Eventually only six proposals were selected, which equaled almost twenty participants from several academic backgrounds (excluding the ToE team).

More and more data from and about governments is available online. Some of these are Open Data and some are Linked Open Data. However, very few of these datasets are connected to each other. In the final Creative Camp the main goal was to contribute to a Web of Data in which these diverse but related datasets are connected. The ToE team therefore invited researchers from various backgrounds to write a proposal for the creation of links between two or more sources of data and come to the third Talk of Europe Creative Camp to carry out their plans. We also strongly encouraged collaboration between researchers with similar or complementary research agendas.

After a plenary session on Monday with introductions to the creative camp, the wider CLARIN framework and the dataset produced by Talk of Europe, the participating teams spent the remainder of the week mostly behind their laptops. On Wednesday, all researchers presented their work-in-progress. Two teams were asked to present their final results of the week during the public symposium on Friday.

The ToE team also invited Markus Strohmaier (professor of Web-Science at the Faculty of Computer Science at the University of Koblenz-Landau and Scientific Director of the Computational Social Science department at GESIS – the Leibniz Institute for the Social Sciences) for the event, who offered valuable insights into measuring political and social phenomena on the web. In his talk ongoing research was presented about gauging political processes – how political actors communicate with voters, how they coordinate with each other, and how they participate in agenda setting and decision making – using data from Twitter during the German Bundestagswahl 2013 and data from the Online Delegative Democracy platform LiquidFeedback.

Subsequently, Wouter van Atteveldt (assistant professor at the VU University of Amsterdam, department of Communication Sciences) gave a presentation on using syntactic information for relational corpus analysis. Maarten Brinkerink concluded the symposium with a short general talk on available Open Culture Data.

Topics CC#3

  1. Wim Peters, Adam Funk and Maarten Marx worked on a follow-up of their project for CC#2. They created a similar functionality as already existing within Polimedia, namely automatically adding high quality links to relevant newspaper articles from the KB newspaper collection to debates in the Dutch Parliament. The team also extracted meaningful key-phrases from the EU and UK proceedings using GATE technology. Finally, they made the ToE dataset searchable from the parliamentary debates repository Search.Politicalmashup.nl.
  2. Alessandro Piscopo, Luis-Daniel Ibáñez, Pavlos Vougiouklis and Nikolaos Liappas linked MEP’s speeches in the EP to tweets and Facebook posts published on their social accounts to concepts, using techniques applied in previous creative camps. Furthermore, they computed values of the sentiment speeches express towards a concept, using state of the art sentiment analysis techniques. They also linked EP speeches to the geographical places mentioned in them, i.e. to IRIs in GeoNames, using state of the art entity recognition techniques.
  3. Mostafa Dehghani, Hosein Azarbonyad and Alex Olieman created an entity linker in which a specialized linker was combined with general entity linking systems to achieve promising precision as well as an acceptable recall. In order to link detected entities to the particular external resources, they also worked on ‘indirect linking’. This entailed links that were created based on the conceptual similarity between documents in the linked corpus and document in the external resource.
  4. Solveiga Inokaitytė, Dainius Jocas, Justina Madravickaitė, Vytautas Mickevičius and Rūta Užupytė linked the debates of the European Parliament with the GDELT Event Database.
  5. Karin van Leeuwen and Bart Vredebregt linked the data of ECJ rulings with proceedings of the EP.
  6. Ewa Kowalczuk linked the EP debates to the Polish Political News dataset.

Concluding remarks
The organizing ToE team thanks the National Library of the Netherlands for their hospitality. We can look back on the third and final camp as a very successful and rewarding event. Again, the contributions of all the participants were of undeniable value for the Talk of Europe project. Each team was asked to write a report on their results achieved during the week: these texts are available here.

Source: Talk of Europe

Posted in Projects, Talk of Europe

Leave a Reply