Social segregation occurs when groups from differing socioeconomic backgrounds living in the same city have little opportunity of interaction with each other. This has profound impacts on socioeconomic outcomes. The measure of social segregation has traditionally been carried out through the use of census records. However, one key issue with relying on this method is that it is dependent on residential data. Increasingly, our social networks extend far beyond where we live, and our actual interactions include where we work, go to school, eat, and spend leisure time.  

As such, we are compelled towards the following question: What other methodology can be explored to measure segregation within a city and improve on the social and urban policy sphere?   

Looking specifically at Singapore, ACI research uses anonymized GPS records, spatially joined to census records, to examine daily movements across geographically refined neighbourhoods. The research concludes that GPS-derived data has the ability to detect segregation by poverty, even with an imperfect proxy and despite having targeted urban policies aimed at social integration. The study uses GPS ping records obtained from CITYDATA.ai, and tests whether real-time movement across neighbourhoods in Singapore correlate with census-derived poverty levels.  

The recent paper titled ‘Segregation Across Neighbourhoods in a Small City’ supports the use of GPS-type data as an emerging tool to understand and study social interactions at a temporal frequency much higher than possible with traditional census-based measures. It is important to understand how this affects the traditional approach towards social and urban policy. Detecting social segregation from GPS-derived movement patterns suggests that there are limits to urban planning policies that specifically aim to socially integrate a diverse group of population. Singapore has a history of policies that aim at social integration such as via public housing allocation along ethnic lines, which itself nests integration by income. In this context, detecting social segregation by real-time exposure of individuals suggests limits to integration policies.  

Past research in this area has already outlined its importance, further enhanced by the new research by ACI. Although the analysis is based on Singapore, the method can be adapted to other metropolitan areas. The efficacy of the GPS-derived data to detect segregation in a city that is meticulous with urban planning and social integration bodes well for the use of GPS data in general. 

By Sunena GUPTA

Researchers: Shu En LEE, Jing Zhi LIM, Lucas SHEN