It is without doubt that social media has become ubiquitous for social networking, content sharing and political branding. However, the content the public shares on these networks remain largely untapped by political hopefuls. In particular, many are not using the predictive powers of social media through sentiment analysis.
Sentiment analysis is a well-studied model in linguistic fields, with different classifiers and language models employed in earlier work of machine-models. The sentiment analysis literature often focuses on analyzing individual documents, or portions thereof (for a review, see Pang and Lee, 2008).
We believe that social media can indeed be used to make quantitative predictions on public sentiment that outperform those of artificial markets. Using the right tools, such as those used by New Politco, a political campaign can use gathered data from social media aggregate the opinions of the collective population and gain useful insights into their behaviour, while predicting future trends in what they would like. Moreover, gathering this sentiment and information on how potential voters converse regarding particular trends, global issues and the like can be helpful when planning upcoming marketing and advertising campaigns. Similarly, previous studies have found that the mere number of messages mentioning a party can reflect the election result. However, it is important to extract the right information when conducting an analysis. The fact that users are discussing political issues online does not mean enough to a politician brand in that they can simply extract meaningful information from the text. What is required is a tool with customizable outcome measures and the ability to detect beyond positive and negative messages. For example, a tweet can contain both positive and negative words, and thus it is important to understand the language being used by the public.
Notwithstanding the fact, that the aforementioned research delivers indication that social media content can be used to authentically forecast political outcomes, we know very little concerning the predictive power of Twitter and other social media platforms for political debates and outcomes. Research does indicate, however, that the sentiment of Twitter messages closely corresponds to political programs, candidate profiles, and suggestion from the media coverage of the campaign trail. With the right analytics, the sentiment profiles of politicians and parties have the possibly to reflect many nuances of the election campaign.