Social segregation research has a long tradition in urban studies. Usually, these studies focus on the residential dimension, using official registries (e.g., census data), which show population distribution at night. Nevertheless, these studies disregard the fact that the population in cities is highly mobile, and its spatial distribution dramatically changes between night and day. The emergence of new data sources (Big Data) creates perfect conditions to consider segregation as a process, by providing the opportunity to dynamically analyse temporal changes in social segregation.

This study uses mobile phone data to analyse changes in social segregation between night and day. Our case study is Medellin (Colombia), a highly socially-segregated, South American city, where social integration policies are being developed, targeting the population in the most disadvantaged neighbourhoods. We use several complementary indicators of social segregation, supplementing them with mobility indicators that help explain changes in spatial segregation between night and day.

The main conclusion is that daily mobility reduces the concentration of a particular group within neighbourhoods and increases the degree of social mixing (exposure) in local settings. This greater social exposure softens local contrasts (outliers) and increases the extension of spatial clusters (positive spatial autocorrelation), so general clustering trends emerge more clearly. The study also makes clear that increased exposure during the day mainly occurs due to the mobility of the low-income population, who are the most likely to leave their neighbourhood during the day and who travel the greatest distances to the most diverse set of destinations.

社会隔离研究在城市研究中有着悠久的传统。通常，这些研究侧重于住宅维度，使用官方登记（例如人口普查数据），显示夜间人口分布。然而，这些研究忽视了城市人口流动性强，其空间分布在白天和黑夜之间发生剧烈变化的事实。新数据源（大数据）的出现为将隔离视为一个过程创造了完美的条件，提供了动态分析社会隔离的时间变化的机会。

本研究使用手机数据来分析白天和黑夜之间社会隔离的变化。我们的案例研究是麦德林（哥伦比亚），这是一个高度社会隔离的南美城市，正在制定社会融合政策，针对最弱势社区的人口。我们使用了几个社会隔离的补充指标，用流动性指标来补充它们，这些指标有助于解释昼夜空间隔离的变化。

主要结论是，日常流动降低了社区内特定群体的集中度，并增加了当地环境中的社会混合（接触）程度。这种更大的社会暴露软化了局部对比（异常值）并增加了空间集群的扩展（正空间自相关），因此总体集群趋势更加清晰。该研究还明确指出，白天暴露的增加主要是由于低收入人群的流动性，他们最有可能在白天离开他们的社区，并且前往最多样化的目的地的距离最远。