Surges of Collective Human Activity Emerge from Simple Pairwise Interactions

Collective human behavior drives a wide range of social, political, and technological phenomena in the modern world. However, while the correlated activity of one or two individuals is partially understood, it remains unclear if and how these simple low-order interactions give rise to the complex large-scale patterns characteristic of human experience. Here we show that a network of email correspondence exhibits surges of collective activity, which cannot be explained by assuming humans act independently. Intuitively, this collective behavior could arise from complicated correlations between large groups of users, or from shared daily and weekly rhythms. Instead, we find that the network is quantitatively described by a maximum entropy model that depends only on simple pairwise interactions, making it equivalent to the Ising model. Remarkably, we find that the learned Ising interactions, which are inferred exclusively from the timing of sent emails, are closely related to the ground-truth topology of email correspondence. Together, these results suggest that patterns of collective activity emerge from simple pairwise correlations, which, in turn, are largely driven by direct inter-human communication.