POLIS Research Seminar Series: "Emoji as emotional proxies in social movement social media"

Postgraduate Researcher Daniel Valdenegro presented this online seminar on 4 November 2020 as part of the School of Politics and International Studies research seminar series. Abstract: In this project my goal is to analyse the evolution of social movements using social media data, identifying the factors that lead to their success, failure and levels of radicalization, using the digital footprint of said social movements, namely tweets or comments on other social media. The main factor to be analysed is the emotional content of the communications associated with these social movements. Emotions are an essential part of the motivational process and as such very important to understanding human and social behaviour. Examples of this are the key role of emotions in motivating participation in activism, or the influence of certain emotions on attitudes towards contingent topics such as immigration, climate change or racism. However, its use in computational social science is less prevalent than its cousin, sentiment analysis, which has the big limitation that most of the lexicons are only available in English, significantly limiting the scope of the analysis. However, social media text may itself have the solution for this problem. Many comments and tweets are already being labelled in terms of their emotional valence using emojis or special characters. This tagging system was born as a way around the limitations of online communication to convey emotional states and more subtle subtext communications. Consequently, these emojis and hashtags can be taken as the emotional label of a piece of text, to be used later to train a machine learning model on the recognition of these emotions. This approach will allow me to study not only English speaking social movements but also Portuguese and Spanish speaking social movements, now common in Latin America. This study collects data from Twitter streaming API, which has a large and very varied international user base. The data was selected using a series of keywords and hashtags related to the social movements of interest in Europe, USA and Latin America.

Valdenegro Daniel

11/4/2020 1:00:29 PM