In order to join virtual sessions, you must be registered and logged-in(Were you registered for the in-person meeting in Denver? if yes, just log in.) 
Note: All session times are in Mountain Daylight Time.

The relational footprints of urban neighborhoods: inferring spatial interactions from social media in Singapore

Authors: Qingqing Chen*, Singapore University of Technology and Design, Ate Poorthuis, Singapore University of Technology and Design
Topics: Geographic Information Science and Systems
Keywords: Human Mobility Pattern, Social Media Data, Survey Data, Home Location, Meaningful Location, Socio-economic, Built Environment
Session Type: Paper
Presentation File: No File Uploaded


Human mobility has been a long-standing interest in Geography and the discipline has applied spatial and relational perspectives to this topic. Traditionally many such approaches use census or survey data that is resource-intensive to collect and often has a limited spatio-temporal scope. The advent of new technologies (e.g. geosocial media platforms etc.) provides new opportunities to overcome these limitations and, if properly treated, can yield more granular insights about human mobility.

In this paper, we use an anonymized Twitter dataset, containing all geotagged messages sent between 2012-2015 in Singapore by ~350k users, to assess this potential. Since tweets have a relatively precise location, we choose to aggregate to a 750m hexagonal grid to circumvent issues with the varying shapes and sizes of administrative regions. We first infer the home location from the spatio-temporal footprint of each user with an ensemble approach for inferring meaningful locations. We then construct a network of connections between the home location and other locations visited within the city, and do so for each user. In aggregated form, this network allows us to analyze the density and direction of urban interactions at a fine spatial scale. To triangulate our findings, we zoom in on three specific neighborhoods and compare the mobility ‘footprint’ derived from the Twitter dataset with a traditional survey conducted in each neighbourhood. We point out that differences in the mobility footprint within and across neighborhoods depend not only on their socio-economic characteristics but also on the structure of the built environment.

To access contact information login