An experienced racial-ethnic diversity dataset in the United States using human mobility data

Keywords: human mobility, urban data science
Abstract Type: Paper Abstract

Abstract

This article details the methods and potential uses for a national experienced racial and ethnic diversity dataset in the United States based on high-resolution mobile phone application data. We calculate measurements of diversity in potential social interactions at the 38.2m X 19.1m scale and 15-minute timeframe for a representative year, aggregated to the Census tract level for purposes of data privacy. We argue that measurements of diversity and segregation, which forms the foundation for equity-based housing and economic development policy, must consider the experienced social context in order to accurately understand the conditions to determine what constitutes as segregation and its outcomes. Given the scope of our data, we also provide validations that show the underlying night-time data to be highly correlated with census residential population and geographically consistent across different degrees of urbanization, and the day-time working population to be highly correlated with work locations. As well, we detail some of the characteristics and limitations of the data for potential use in national, comparative studies.