An extended spatiotemporal exposure index for urban racial segregation

The Segregation Index quantifies the degree of segregation of social groups or classes. Because of the increasing use of fine-grained spatiotemporal activity and flow data, the conventional segregation measurements’ inclusiveness is challenged. We add population flow to the conventional place-based spatial exposure index to identify spatiotemporal segregation changes. Specifically, we considered the population-flow network, hierarchical structure, and time. In Chicago’s demonstration case study, we first used the time-dependent Twitter Origin-Destination flow matrices and their hierarchical structure information to estimate interactions between areal units at the neighborhood level. Then we computed the new population composition of units based on their interactions with other units and estimated the proposed spatiotemporal exposure index for different times. Finally, we systematically compared their differences with the conventional indices at global and local scales to see how population-flow patterns affect the exposure index. The results show that the population-flow patterns reflect valuable information in neighborhood interactions in temporal and spatial dimensions, but it is missing information in the conventional segregation computations. Furthermore, we emphasize that the hierarchical structures of flow patterns and the choice of appropriate parameters are also important factors for a rational segregation evaluation.

隔离指数（Segregation Index）用于量化社会群体或阶层的隔离程度。随着细粒度时空活动与流动数据的应用日益广泛，传统隔离测度方法的包容性受到了挑战。我们将人口流动纳入传统基于地域的空间暴露指数（spatial exposure index），以识别时空隔离的变化。具体而言，我们纳入了人口流动网络、层级结构与时间维度。在芝加哥的示范案例研究中，我们首先采用随时间变化的推特（Twitter）起讫点（Origin-Destination）流动矩阵及其层级结构信息，估算街区尺度下区域单元之间的互动关系。随后，我们基于区域单元与其他单元的互动关系，计算出各单元的新人口构成，并针对不同时段估算本文提出的时空暴露指数。最后，我们在全局与局域尺度下，系统对比了新指数与传统指数的差异，以探究人口流动模式对暴露指数的影响机制。研究结果表明，人口流动模式能够反映时空维度下街区互动的宝贵信息，而这部分信息在传统隔离计算中是缺失的。此外，我们强调，流动模式的层级结构与合适参数的选取，同样是开展合理隔离评估的重要影响因素。