Human activities---from voter mobilization to political protests---increasingly take place in online environments, providing novel opportunities for relating individual behaviours to population-level outcomes. The recent availability of data sets that capture the behaviour of individuals participating in online social systems has driven the emerging field of computational social science, as large-scale empirical data sets enable the development of detailed computational models of individual and collective behaviour. Given the inherent limitations of observational data, it is crucial to investigate the extent to which models of collective dynamics can distinguish between different individual-level mechanisms. Here we introduce a simple generative model for the collective behaviour of millions of social networking site users who are deciding between different software applications. Our model incorporates two distinct components: one is associated with recent decisions of users, and the other reflects the cumulative popularity of each application. Importantly, although various combinations of the two mechanisms yield long-time behaviour that is consistent with data, only models that strongly emphasize recent popularity of applications over their cumulative popularity reproduce the observed temporal dynamics. Our approach demonstrates the value of even very simple generative models in understanding collective social behaviour, and it highlights the need to address temporal dynamics---not just long-time behaviour---when modelling complex social systems.
A Simple Generative Model of Collective Online Behaviour
James P. Gleeson, Davide Cellai, Jukka-Pekka Onnela, Mason A. Porter, Felix Reed-Tsochas