The impact of residential segregation on socio-spatial equity and urban stability is a crucial concern for sustainable urban development. Advances in geographic data, such as mobile signaling data (MSD), provide new opportunities to understand residential segregation in the spatiotemporal context at the traffic analysis zone (TAZ) level. Using recordings of 19.3 million mobile users in Shanghai, we measured the diversity of spatiotemporal behavior among residents living in central urban (RC) and noncentral urban areas (RN). We eliminated the influence of different population densities by the odds ratio (OR). A SARIMA model was formulated to forecast the OR in the central and noncentral urban. The findings suggest that residential segregation can be inferred through the spatial distribution and changes of residents’ activity space. TAZs with high homogeneity levels between RC and RN in central urban declined 83.7 % on the weekend than workdays, while in the noncentral urban increased 149.3 %, which reflected employment−housing spatial mismatch of RN and potential less satisfaction of leisure activities of RC. The periods 12:00−14:00 and 18:00−20:00 were confirmed as having the highest inter-urban mobility in the central and noncentral urban, respectively. The research contributes to understanding the social welfare and sustainable development of Chinese cities.