The mode, timing and frequency of travel by residents of urban neighbourhoods is affected by sociodemographic and environmental factors. Among these, the layout and connectivity of the street network is amenable to design by urban planners and developers. Here we focus on street connectivity as a variable influencing mobility in cities by examining GPS-enabled mobile phone location data volunteered by a group of 127 university students (234,709 h of behavioural observation over a six-year period). We used a Bayesian beta regression framework to model the proportion of time spent inactive, walking and travelling at vehicle speeds, relative to street connectivity and other environmental attributes measured within a radius of home. Results indicated that lower street connectivity, measured using a simple measure we call network warp, is associated with more time spent inactive and more time travelling at vehicle speeds by students. The proportion of time spent walking was higher in areas with more street connectivity, and for student homes that were closer to campus. Our study confirms the importance of street connectivity as a factor influencing the walkability of neighborhoods and the selection of passive forms of transport, and builds on earlier studies of this relationship by incorporating longitudinal data with high spatial and temporal resolution. We conclude that crowdsourcing data that is recorded automatically by GPS-enabled mobile phones can provide an accessible and flexible evidence base to support the design of urban areas.