Between the 23th and 26th May 2019 millions of people voted for the elections of the European Parliament. On the wave of the political changes that are 'reshaping' the European Union, with one of its historical members about to leave it (Newsflash: Brexit), the elections have been a significant political moment.
Europinion aims at identifying people’s reaction to the EU elections, in terms of either positive or negative opinion, as it emerged in the very aftermath of the vote. We decided to apply sentiment analysis to Twitter, since we thought that it could have been a rich and appropriate source of data for our project, which aims at mapping opinions about a pan-european political matter.
Aa a criterion to select the tweets we searched for the expression “European elections” translated into all the EU official languages in Twitter. Given that the final results of the elections were published on Monday, the persists of tweets tweeted after the 27th of May. Due to scarcity of data for some of the countries we abandoned the idea of creating a proportion between the selection of the number of tweets per country and the number of seats that country is assigned for the European Parliament. Besides, our research was driven by languages rather than countries. However, it can be noticed that generally the higher the number of seats, the higher the number of tweets tweeted in the language spoken in that country. (check our Data). Overall, we collected approximately 100,000 tweets.
Twitter is the social network used to discuss current issues, especially political ones. Its use has spread all over the world. Twitter’s users can express their agreement, disagreement or complaint about the ongoing discussions.
Our ambition was to represent the sentiment of all EU citizens. We tried to achieve this objective by acquiring data for all languages which the EU officially supports: From English to Maltese. (List of Languages)
Sentilo assigns an average positive or negative score to each tweet, sometimes both to the same tweet. It also assigns a positive or negative score to some keywords it identifies in the tweet.
This Tweet was the most liked and retweeted Tweet concerning the European Election in the period 27th May to 2nd June.
createdBy: JuliaHB1 – hasLocation: London – hasDate: 2019-05-27 – hasTime: 06:55:20
hasRetweet: 7078 – isFavorite: 19274 – hasAvgNegative: 0.5456 – hasAvgPositive: 0.0378
hasText: Leave wins 52% in the EU referendum in 2016. Remoaners: but not by enough. Leave wins 80%+ the 2017 general election. Remoaners: but it wasn’t clear. Leave wins 2019 European elections. Remoaners: but the combined vote of the Remain losers is higher. Anyone seeing a pattern here?