Social Media as a Lens on Collective Emotion

During the XX April International Academic Conference on Economic and Social Development, scheduled this year for April 9-12 at the Higher School of Economics, Dr David Garcia of the Complexity Science Hub Vienna and the Medical University of Vienna, Austria will present a report entitled ‘The digital traces of collective emotion’ at a session on ‘The Wellbeing of Children and Youth in the Digital Age.’ Ahead of the conference, Dr Garcia spoke with the HSE News Service about his conference presentation, his research, and plans for ongoing collaboration with HSE colleagues.

— At this year’s April Conference, you will represent Complexity Science Hub Vienna and Medical University of Vienna. What are the main areas of focus for these two institutions? 

— One of the main aims of the Complexity Science Hub is to make sense of Big Data, and with the Medical University of Vienna we try to understand health and well-being from a complex systems perspective. We analyse large-scale datasets and apply computational models and complexity science methods to understand well-being from a holistic perspective that takes into account the environment, social interactions, and structural aspects of heath care and social systems.

— What are some of the main themes of the report you’ll be presenting at the conference? 

— I will introduce an overview of the opportunities and limitations of analysing social media data for the social sciences, which is commonly known as Social Data Science or Computational Social Science. I will illustrate what we can learn from digital traces with a recent study we did on the collective emotions after the Paris terrorist attacks of November 2015, a phenomenon we were able to analyse thanks to the millions of public tweets that French people posted after the attacks.

— Your recent research has focused on understanding emotions and social interactions in the digital society. Can you share some of your main findings with us? 

— We found that the terrorist attacks triggered a collective emotion in which people shared emotions with each other through Twitter, thus having a longer emotional experience than if they were isolated. We then analysed the long-term behaviour of people in France and found that those who were involved in the collective emotion were showing higher levels of language associated with solidarity and prosocial behaviour. We interpret this as an example of social resilience in which a community increases its solidarity through collective emotions after a trauma. We also have further research on testing theories from sociology and psychology using social media data, which often allows us to complement traditional methods in the social sciences.

— How has digitalization of society caused people to change? 

— Research is finding certain ways in which information and communications technology (ICT) has affected human behaviour. For example, we have shorter attention spans, and text and news in general are becoming briefer and targeted towards social sharing. Certain activities barely happen without mediation of the internet, for example job seeking. Even word meanings seem to have adapted to the way we type them on the keyboard.

Overall, however, we remain the same with respect to most things we do and the most fundamental building blocks of society. What digitalization brings is a new way to quantify and analyse human behaviour at completely new scales and resolutions.

— How did your collaboration with HSE start, and what are some of the specific joint projects you’re working on?

— I work with Ivan Smirnov and Lisa Sivak on the analysis of wellbeing and emotions through the digital traces left on the Russian social network VK. Our collaboration was the initiative of Ivan – he came to visit us last summer to work together on these topics. This is one in a longer list of collaborations between him and people at the Medical University of Vienna, for example Professor Stefan Thurner.

Our research on VK is still a bit preliminary, so we don't yet have many final results we can promote right now. I can say that what we are doing is analysing emotional expression in public VK posts with the aim of calibrating a model of wellbeing and against self-reported data.