This event will be a HYBRID event. Please register to join this seminar. You will then receive an email with the dial in details for those attending virtually or further instructions for those attending in person. Please check your spam/junk folders.
The event will be followed by tea/coffee & biscuits in the top floor atrium space outside the INET offices for those attending in person.
Identifying intervention points in socioeconomic systems is key to advance climate change mitigation. To that end, Twitter has been extensively used to analyse climate change’s public discourse and proven to be a successful tool for monitoring public perception in other scenarios. Here we study the Twitter conversation on climate change during 2019, when important climate-related movements occurred, such as “Extinction Rebellion” or “Fridays For Future”. Our goal is to characterise the complexity of the overall conversation on Twitter by studying the impact of its leading users on their audiences and analysing how these audiences interact.
We find that a few leading users dominate the Twitter conversation on climate change, with a massive inequality of the users' impact characterised by an average Gini index of 0.89. The dynamics of the most persistent leading users show that there are groups of users with contrasting ideologies, which we identify as climate activists and climate skeptics.
Moreover, we quantify who the leading users are talking to. Thus, we define the audience of user i as the set of users that retweeted her and her chamber as the users retweeted by her audience. The chamber gives us a proxy of the information sources of i's audience. We constructed similarity networks of leading users weighted by the overlap of their chambers, where we found a clear community structure that matches the users ideology, showing high levels of polarization. This work shows that by studying a small set of users, we can tell a great deal about thedynamics of the Twitter conversation. We will further analyse the network structure ofthe chambers and focus on the dynamic flow of information between them, as well as the biases induced in the conversation (eg. bots, impact of users in chambers).