Hands-on session by: Laura Hollink
Title: Talk of Europe
Abstract: In this hands-on session we will explore the data of the European Parliament. The Talk of Europe project has translated the debates of the European Parliament to Linked Open Data (LOD), and linked it to a variety of external data sources. We will use the LOD query language SPARQL to get both qualitative and quantitative results, and the statistical software package R to visualize the results. First, I will show how to run SPARQL queries on this data that combine full txt search through the transcripts of the debates with structural properties such as demographics, memberships, and roles of speakers. Starting very simple, we will build up to increasingly complex queries that allow a researcher to zoom in to exactly that part of the data that is of interest to her. Second, we will use R to process and visualize the results. Finally, using the above mentioned approaches, we will explore and critically assess a part of the data that is still under development: topics that were automatically detected in the spoken words using the JEX software of the Joint Research Center of the European Commission. The flexibility of SPARQL allow us to inspect and compare relevant subsets - e.g. topics detected in English or Bulgarian speeches, or topics detected in translations from English to Bulgarian - in order to asses the effect of languages and translations on the topic detection tool.
Back to the Workshop page.