Extracting spatiotemporal commuting patterns from public transit data

• With accelerated migration rates, cities are transforming considerably and trying to keep up with the changing mobility needs of its citizens. Transport demand analysis heavily relies on census information or modelling based complete trajectories individuals; data that gets quickly outdated. We propose use tractable privacy-preserving develop a framework for transportation over time. apply our Greater London region consisting 4 million traces mobility. find individual profiles can be easily extracted from such simple day’s mobility, which also reveals structure urban areas. This evaluation transit system lot about efficiency, mixed-use potential developing areas safety sustainable growth. Public networks in crucial addressing citizens work, services leisure. The rapid changes demographics pose several challenges efficient management services. To forecast demand, planners often resort sociological investigations, population either difficult obtain, inaccurate How we then estimate variable mobility? method identify spatiotemporal public city. Using Gaussian mixture model, decompose empirical ridership into set temporal representative any given day. A case ≈ 4.6 daily primary mode underground distinct commuting profiles. weighted these generate station traffic remarkably well, uncovering spatially concentric clusters needs. Our results suggest used stations exhibit patterns generally located cluster central business district away centre city largely single residential Overall, identifying mixed spatial diverging macro indicates approach may useful detailed understanding integrated planning heterogeneous travellers.