CIVICA Research Blog

Twitter reveals our emotions throughout the pandemic

Policy interventions during the COVID pandemic have the power to spur emotions, researchers learned in the early stage of "EmoMap - Mapping Emotions during COVID-19 pandemic using Twitter data," a CIVICA Research project carried out by Bocconi University, Hertie School, and Central European University. Policymakers, though, won’t be happy to discover that the predominant emotion connected to politics in this period is anger.

“We are using Natural Language Processing (NLP), a machine learning technique, to analyse millions of tweets collected during the pandemic in order to investigate the formation and evolution of emotions,” said Nicoletta Balbo, a sociologist at Bocconi and Principal Investigator of the project.

This project aims to establish a general tool for mapping emotions from social media activity. However, the COVID-19 pandemic provides a unique natural experiment because it has been an amplifier of emotions and fears, and an accelerator of the digitalisation process.

The corpus of tweets currently under analysis come from Italy, but a similar process will be repeated for Germany and Hungary.

In the near future, a publicly accessible dashboard showing the real-time evolution of citizens' emotions, perceptions and opinions since the beginning of the pandemic will be available on Bocconi’s Covidcrisislab website.

In its current version, still under construction, the dashboard allows users to understand how emotions vary by groups and over time. Segmentation is possible by geographical area (Italian regions), gender, and age in relation to four emotions (anger, fear, joy, and sadness) and four discussion topics (politics, vaccine, economics, and health). Users can also split the topic-emotion relationship by age group, choosing, for example, to observe the evolution of tweets by people in the 19-29 age bracket expressing fear in relation to the economy.

For example, the dashboard reveals that, from February 2020 to February 2021, anger was the most common emotion in tweets related to politics, while fear prevailed in those related to the vaccine. The relative majority of tweets regarding health and economics are neutral. “Changes in emotions are clearly observed around the announcements of new COVID regulations,” Professor Balbo said.

The Northern Italian regions most affected by the pandemic in its first year (and notably Lombardy) have the lowest share of neutral tweets related to all the topics.

The EmoMap project is currently moving on to a successive, interpretative stage. “Now we must – and thanks to the dashboard, we can – understand the way beliefs form and vary by groups and over time; how trust in government, public health authorities, experts, and in the future has evolved during the pandemic; the social structure underlying patterns and diffusion of emotions and opinions; and to what extent emotions and perceptions can be linked to life course choices of individuals,” Prof. Balbo said. “And furthermore, discussions with the network of scholars formed around this project are stimulating ideas for new research.”

Participants in the project also include demographers (Arnstein Aassve and Letizia Mencarini), epidemiologists (Alessia Melegaro) and data scientists (Dirk Hovy) from Bocconi University, along with data scientists from the Hertie School (Slava Jankin and Hannah Bechara) and Central European University (Márton Karsai, Júlia Koltai, and Adriana Manna).

CIVICA Research brings together researchers from eight leading European universities in the social sciences to contribute knowledge and solutions to the world’s most pressing challenges. The project aims to strengthen the research & innovation pillar of the European University alliance CIVICA. CIVICA Research is co-funded by the EU’s Horizon 2020 research and innovation programme. To stay up to date with CIVICA Research developments and opportunities, subscribe to the newsletter.

Written by Fabio Todesco (Bocconi University)


Leave a reply

×

Name is required!

Enter valid name

Valid email is required!

Enter valid email address

Comment is required!

* These fields are required.

Be the first to comment