Unveiling epidemic dynamics: harnessing the synergy of social media data and mobility patterns during COVID-19

The convergence of social media data and mobility patterns presents a unique opportunity to delve deeper into societal behavior and its implications on epidemic dynamics. Social media platforms have become veritable repositories of real-time, user-generated data, reflecting public sentiments, discussions, and perceptions regarding health concerns, including infectious diseases. Concurrently, mobility patterns, derived from transportation usage, geolocation services, and movement data, offer insights into population movements and travel behaviors critical for understanding disease spread. This paper proposes an approach exploiting social media data and mobility patterns to perform epidemic predictions, whose experimental evaluation has been carried out on COVID-19 pandemic data. Leveraging these diverse data sources, we aim to uncover synergies and correlations between online discussions and travel behaviors that can contribute to more accurate and proactive epidemic forecasts. Our central focus lies in discerning the alignment between peaks in social media discussions and corresponding fluctuations in mobility patterns. By identifying and analyzing these alignments, our aim is to clarify their potential as predictive indicators for upcoming epidemic trends. Results obtained from real-world datasets about the city of Chicago (USA) demonstrate the efficacy of the proposed method in predicting the spread of the epidemic accurately. The explainability analysis reveals a significant correlation between tweet content and actual COVID-19 data, affirming Twitter’s credibility as a dependable indicator of epidemic spread. This underscores the growing importance of social media user-generated data as a valuable resource for monitoring and comprehending epidemic outbreaks.