People-based Segregation Indices: Measuring Segregation with Individual’s Activity Space and Demographics in U.S. 50 Largest Cities Using Social Media

Keywords: people-based segregation indices, segregation measurement, racial-ethnic segregation, human mobility, social media
Session Type: Virtual Paper Abstract
Day: Friday
Session Start / End Time: 2/25/2022 02:00 PM (Eastern Time (US & Canada)) - 2/25/2022 03:20 PM (Eastern Time (US & Canada))
Room: Virtual 26

Abstract

The five-dimension of segregation (i.e., evenness, exposure, concentration, centralization, and clustering) and their variants have been challenged for decades about their validity on measuring segregation with place-based demographics, which raises the issue of the modifiable areal unit problem (MAUP) and limits the segregation scope from various socioeconomic spaces to a single space (e.g., residential). Another deficiency of using place-based demographics is that most studies only focused on spatially separated groups and neglected the fact that segregation could also induce segregated facilities and services with unequal accessibilities among different groups, and therefore the coverage of existing dimensions is narrowed in terms of representing social inequality. In response, we created a people-based conceptual framework with refined dimensions (“3E”– extent, environment and exposure) based on individual’s activity space and demographics, aiming to comprehensively reveal social inequality due to segregation. With individual information inferred from long-term Twitter data collected from Dec 2013 to May 2015, we also proposed the people-based segregation indices, which sufficiently addressed the issues caused by using place-based demographics and revealed novel findings of urban racial-ethnic segregation across the U.S. 50 largest cities. The preliminary results indicate that transit rate is negatively correlated to economic isolation and the same between land use mix rate and travel distance dissimilarity. Additionally, cities with a larger distance dissimilarity may also suffer a worse economically-isolated environment. Lastly, the comparisons between the people-based results and previous place-based ones in Chicago and Miami demonstrate that our people-based indices would be more comprehensive and advanced for segregation measurement.