Election Forecasts With Twitter

Abstract

This study investigates whether microblogging messages on Twitter validly mirror the political landscape off-line and can be used to predict election results. In the context of the 2009 German federal election, we conducted a sentiment analysis of over 100,000 messages containing a reference to either a political party or a politician. Our results show that Twitter is used extensively for political deliberation and that the mere number of party mentions accurately reflects the election result. The tweets' sentiment (e.g., positive and negative emotions associated with a politician) corresponds closely to voters' political preferences. In addition, party sentiment profiles reflect the similarity of political positions between parties. We derive suggestions for further research and discuss the use of microblogging services to aggregate dispersed information.

Using Twitter Sentiment to Forecast the 2013 Pakistani Election and the 2014 Indian Election

“…There is a lively debate on this topic by others who have (prior to us) tried to forecast the winners of elections with Twitter. 5,6 That this can be done cheaply and accurately is conclusively demonstrated by this article-but whether it can be done consistently for multiple countries and multiple elections requires further proof. For this, a head-to-head controlled experimental comparison is necessary.…”

Twitter and Television in a Hybrid Media System

“…Additionally, the hypothesis about social media being indicators of electoral behaviors is not new; it has been advanced in other contexts as well. In fact, some scholars argue that through the Twittersphere, it is possible to identify indicators that clearly describe political opinions (Tumasjan et al 2011) or to replace traditional polling (O'Connor et al 2010). However, these hypotheses have already been refuted (Gayo-Avello, Metaxas, and Mustafaraj 2011), and in the Italian Twittersphere, we found important evidence that supports the findings of Gayo-Avello, Metaxas, and Mustafaraj.…”

Twitter in academic events: A study of temporal usage, communication, sentimental and topical patterns in 16 Computer Science conferences

“…According to the cur-3 rent statistics as published by the company itself 1 , there are more 4 than 300 million monthly active users, generating over 500 million 5 tweets every day, expressing their opinions on several issues all over 6 the world. This massive amount of data generated by the crowd has 7 been shown to be very useful in many ways, e.g., to predict the stock 8 market [8], to support government efforts in cases of natural disasters 9 [1], to assess political polarization in the public [11,41], and so on. sample of scientific events at the same time, which would help to 29 generalize the results obtained.…”

“…For instance, by investigating us-37 age patterns or content published, conference organizers could be 38 able to better understand their attendees. By identifying (i) which 39 are the most popular conference topics, (ii) who emerge as the online 40 community leaders, (iii) how participants interact, (iv) what is the 41 users' sentiment towards the event, etc. organizers and researchers 42 can create attendance and participation prediction models, they can 43 also quantify trending topics in their respective research community, 44 or they might invent novel recommendation systems helping aca- 45 demics connect to each other with greater ease [17, 46,47].…”