Social media have become a prominent source of data for researchers interested in the analysis of social interactions and collective dynamics. This interest responds, in part, to the proliferation in recent years of massive events coordinated through online networks. Political demonstrations, protests, riots, and shifts in public opinion respond to the coordinating potential of communication technologies – which are now pervasive and constantly pulsating with information. This talk will consider recent research examining how online networks mediate those collective dynamics, with a focus on the structural and temporal signatures that characterize the emergence of massive mobilizations. The goal is to encourage discussion around two interrelated questions: how do recent developments in data science and computational methods help us deal with the problem of scale? And do those methods allow us to use social media not just as a measurement device but also as an early warning system? The talk will argue that finding a solution to these problems requires a genuine (and fearless) collaboration across traditional disciplinary divides.