Using AI to control energy for indoor agriculture
30 September 2024
Published online 6 April 2016
Analysis reveals changing patterns in social media in the run-up to major events.
An international team of researchers has identified patterns in social media leading up to major events, opening the door to understanding what drives these collective behaviours.
Together with colleagues in Spain, France, and the USA, Javier Borge-Holthoefer of Qatar’s Hamad Bin Khalifa University analysed data from Twitter before, during, and after several major events, such as significant protests or media announcements. Their analysis revealed signatures associated with events which were driven by an internal momentum rather than an external trigger1.
Twitter conversations became more rapid leading up to an event, even though the overall volume of communication remained low. “We are aware that the early increase in information exchange might raise expectations about forecasting”, says lead author Borge-Holthoefer, cautioning that the technique’s predictive capacity is limited by the need to identify relevant conversations for analysis before an event.
After each event, the flow of information switched from a hierarchy dominated by a few sites to a more distributed pattern.
“Online social platforms flatten — to a certain extent — the capacity to influence”, explains Borge-Holthoefer. “This is in stark contrast to how things happened during the US Civil Rights movement in the [19]60s, for example, which has been a benchmark for studies of collective action. In a sense, our work aims to update those assumptions.”
The new method and the patterns it revealed will help researchers studying the mechanisms driving collective action.
doi:10.1038/nmiddleeast.2016.41
Borge-Holthoefer, J. et al. The dynamics of information-driven coordination phenomena: a transfer entropy analysis. Sci. Adv. 2, e1501158 (2016).
Stay connected: