Ad verba per numeros
Brief comment on a paper: "Pacts with Twitter. Predicting voters' indecision and preferences for coalitions in multiparty systems"
A few days ago Frederic Guerrero-Solé (@guerrerosolef) pointed me to a paper he had recently published. The abstract is very explicit so I'll provide it verbatim:
Wednesday, June 4, 2014, 01:46 PM
Wednesday, June 4, 2014, 01:46 PM
In recent years, social networking sites (SNS) have become enormously popular, in particular microblogging sites such as Twitter. Twitter is nowadays one of the most used SNS for social, economic and political communication. Drawing on well-known characteristics of social networks and human behavior, i.e. the homophilic behavior of individuals, the power law distribution of influence and retweeting, and the nature of endorsement and the reduction of noise of retweeting, this article proposes a two-step method to first uncover the structure of the network of the top influential Twitter users in a political discussion and second based on the resulting structure of political clusters, predict the voters perception of the closeness between parties, the proportion of undecided voters between two given parties and the support for pacts between parties. The method analyzes the overlaps of communities of retweeters of the most influent users in a political conversation, and applies it to the Catalan elections in 2012. Comparing our results with those of the Spanish Center for Sociological Research, we show that the overlaps between parties' communities of retweeters are a good predictor of voters' indecision and preferences for post-electoral parliamentary support and coalitions.
In short, Frederic et al. have analyzed political discussion by exploiting retweets in a quite novel way; instead of studying the RTing cascades they have paid attention to the communities that co-RTin induces. Furthermore, they have analyzed a multiparty scenario which, IMHO, is much more interesting than common analyses of Democrats vs Republicans that pervade the literature.Needless to say, the paper has limitations as any other political research conducted on Twitter data but I hope to see this approach replicated in other multiparty scenarios in the near future.As a bonus track I provide a number of additional references that can be of interest for anyone working in the area, and/or trying to extend the work by Guerrero-Solé et al.
- First, a couple of papers providing support to the idea that tweets convey (in a latent and subtle way) voting intention: DiGrazia et al., 2013; Jensen & Anstead, 2013).
- Then, a couple of papers discussing the worrying bias caused by relying on samples of tweets: González-Bailón et al., 2012; Morstatter et al., 2013.
- A nice discussion about the meaning of edited retweets regarding political discussion: Mustafaraj & Metaxas, 2011.
- And, finally, some recent literature about community detection: Fortunato, 2010; Xie et al., 2013; Yang & Leskovec, 2013.
- DiGrazia, J., McKelvey, K., Bollen, J., & Rojas, F. (2013). More tweets, more votes: Social media as a quantitative indicator of political behavior. PloS one, 8(11), e79449.
- Fortunato, S. (2010). Community detection in graphs. Physics Reports, 486(3), 75-174.
- González-Bailón, S., Wang, N., Rivero, A., Borge-Holthoefer, J., & Moreno, Y. (2012). Assessing the bias in communication networks sampled from Twitter. arXiv preprint arXiv:1212.1684.
- Jensen, M. J., & Anstead, N. (2013). Psephological investigations: Tweets, votes, and unknown unknowns in the republican nomination process. Policy & Internet, 5(2), 161-182.
- Morstatter, F., Pfeffer, J., Liu, H., & Carley, K. M. (2013). Is the sample good enough? Comparing data from Twitters streaming API with Twitters firehose. Proceedings of ICWSM.
- Mustafaraj, E., & Metaxas, P. T. (2011, August). What Edited Retweets Reveal about Online Political Discourse. In Analyzing Microtext.
- Xie, J., Kelley, S., & Szymanski, B. K. (2013). Overlapping community detection in networks: The state-of-the-art and comparative study. ACM Computing Surveys (CSUR), 45(4), 43.
- Yang, J., & Leskovec, J. (2013, February). Overlapping community detection at scale: a nonnegative matrix factorization approach. In Proceedings of the sixth ACM international conference on Web search and data mining (pp. 587-596). ACM.