Data and Methodology in the Twitter EP2019 Analysis

The chapter introduces the data collection process and methods used in the study. The main dataset was assembled form material collected from seven EU countries that represented so-called Twitter countries during the 2019 EP elections: the Netherlands, Germany, Finland, Italy, Spain, Ireland and the UK. The countries cover the South-North and Centre-Periphery dimensions in Europe and adequately follow the system models of politics and media devised by Hallin and Mancini. The data were gathered in real time during the EP election campaign in May 2019, based on hashtags. Two datasets were collected: raw data comprising 1,552,674 tweets from 222,169 accounts from all 27 EU countries covering all actors participating the campaign discussions, and a more selective main database of 49,492 tweets belonging to 2512 politically affiliated accounts in the seven above-mentioned countries. The raw data were used in computational topic modelling to find the timeline of various topics, and how they relate to each other. The computational and manual word frequency analysis of the main data was used to figure the themes favoured by various political actors in specific countries, and a network analysis was carried out to map the activities of these tweeters and their relationships. In addition, the chapter shows the methodological particularities in each country and discusses the 2019 EP elections as a specific context for the study.