Transport networks (maritime, road, telecommunications networks...) share a common structure: nodes connected by edges whose topology governs how flows circulate or are interrupted. Large-scale data and tools from network science and graph signal processing now allow rigorous study of traffic on these infrastructures, yet disruption dynamics remain poorly understood. This thesis focuses on disruptions impacting these networks, with particular attention to disruptions caused by social movements, a subject studied extensively in social and political science but rarely quantified in network science. Rather than modeling the movements themselves, we study their impact on the networks they disrupt. We introduce a general formalism that turns mobility data into temporal networks and develop a framework based on differentiation and clustering to characterize these effects quantitatively and unveil their temporal and structural extent.