Topics evolution through multilayer networks; Analysing 2M tweets from 2022 Qatar FIFA World Cup
Andrea Russo, Vincenzo Miracula, Antonio Picone
TL;DR
This study addresses how topics evolve on social media during a large-scale event, using a multilayer network where tournament stages form temporal layers. It collects approximately 1,923,283 tweets (with ~2.2 million total) and builds a Bigram-based word network with a 300-word cap per layer, analyzed via Gephi and modularity-based community detection. Results show a shift from Group stage discussions centered on crime and BTS to later rounds focused on teams, players, and the Messi–Mbappé final narrative, with the network resulting in 858 nodes and 1,041 edges. The work demonstrates a contextual, layered approach to visualizing evolving social events and highlights both the methodological benefits and challenges of cross-layer analysis for uncovering emergent social dynamics.
Abstract
In this study, we conducted a comprehensive data collection on the 2022 Qatar FIFA World Cup event and used a multilayer network approach to visualize the main topics, while considering their context and meaning relationships. We structured the data into layers that corresponded with the stages of the tournament and utilized Gephi software to generate the multilayer networks. Our visualizations displayed both the relationships between topics and words, showing the word-context relationship, as well as the dynamics and changes over time by layer of the most frequently discussed topics.