.svg)
Mobility data for urban planning research
Mobility data reconstructs trips from location traces left by smartphones and connected vehicles, recording where a device dwelled, for how long, and which places it visited in what order. That yields an origin destination picture refreshed daily, in place of a household travel survey fielded once a decade to a few thousand respondents in one metropolitan area. It makes it possible to ask who actually reaches a station, a park or a library, which neighbourhoods they came from, and which groups never arrive at all. The same continuity supports before and after measurement around a service change, a road closure or a storm, at tract and block group grain.
How urban planners use mobility data
Transit access, ridership and unmet demand
Planners pair station boarding counts with device trajectories to estimate not only how many trips a corridor carries but how many plausible trips are never made, comparing observed origin destination flows against the flows that residential and workplace locations would predict. The design turns on the share of a neighbourhood's devices that appear at a station or along a route at all, which separates low ridership caused by thin service from low ridership caused by thin population, and it supports equity comparisons by home block group demographics. Mode is inferred rather than observed, so a bus trip and a car trip along the same arterial are distinguished only by speed and stop pattern, and short walking access legs are frequently missed.
Green space, parks and who actually arrives
Distance to the nearest park says nothing about whether anyone goes, so planners combine visit counts at park polygons with the home census block group of visiting devices to measure the catchment a green space really serves and how that catchment differs from the population living within walking distance. This supports designs that test whether a physical barrier such as an arterial road, a rail corridor or a river suppresses visits from one side, and whether new amenities change who visits rather than only how many. Panel composition varies across neighbourhoods, with older and lower income residents typically underrepresented, so visit shares need weighting to home area population before differences between groups are read as differences in behaviour.
Disaster disruption and the recovery of everyday movement
Because the panel is continuous, planners take the weeks before a flood, storm or extreme heat event as a baseline and measure how far trip counts, trip distances and visits to essential destinations such as pharmacies, grocery stores and shelters depart from the same day of the week, then how long they take to return. The variable is a deviation series rather than a level, which makes urban, suburban and rural counties comparable even when their device densities differ by an order of magnitude. Devices that evacuate, lose power or lose network coverage leave the panel entirely, and that is the population most disrupted, so displacement and data loss are difficult to separate at the peak of an event.
Mobility datasets available on Dewey
Veraset
Device level movement records with latitude, longitude and timestamp, which planners use to build their own origin destination matrices, catchment polygons and dwell time measures rather than inherit someone else's aggregation.
View datasetAdvan
Visit counts and visitor home census block groups for millions of US points of interest, suited to questions where the unit of analysis is a destination such as a park, a library or a station area and the point is who travels to it.
View datasetSafeGraph
Point of interest geometry, attributes and visit patterns, giving planners a consistent place base for defining what counts as a park, a transit station area or a grocery store before any mobility measure is attached to it.
View datasetMore published work using mobility data
Bartosz Bonczak, Shlomo Angel & Constantine E. Kontokosta · Nature Cities · 2026
Huan Ning1, Zhenlong Li, Manzhu Yu, Xiao Huang, Shiyan Zhang, Shan Qiao · Working Paper · 2026
Sarala Dhakal, Ruijie Bian, Tara Tolford · Journal of Transportation Engineering, Part A: Systems · 2026
.svg)