Social and spatial ethnic segregation: a framework for analyzing segregation with large-scale spatial network data

While ethnic segregation plays an important role in determining the development trajectories of many countries, empirical measures of the dynamics of segregation remain rudimentary. In this paper, we develop a new computational framework to model and measure fine-grained patterns of segregation from novel sources of large-scale digital data. This framework improves upon prior work by providing a method for decomposing segregation into two types that previous work has been unable to separate: social segregation, as observed in interactions between people, and spatial segregation, as determined by the co-presence of individuals in physical locations. Our primary contribution is thus to develop a set of computational and quantitative methods that can be used to study segregation using generic spatial network data. A secondary contribution is to discuss in detail the strengths, weaknesses, and implications of this approach for studying segregation in developing countries, where ethnic divisions are common but data on segregation is often plagued by issues of bias and error. Finally, to demonstrate how this framework can be used in practice, and to illustrate the differences between social and spatial segregation, we run a series of diagnostic tests using data from a single city in a large developing country in South Asia. The case study we develop is based on anonymized data from a mobile phone network, but the framework can generalize easily to a broad class of spatial network data from sources such as Twitter, social media, and networked sensors.