Multi-Scale Spatio-Temporal Evolution Characteristics and Influencing Factors of Jobs-Housing Balance in Shenzhen in the Context of COVID-19 Pandemic

The outbreak of the coronavirus disease (COVID-19) pandemic, subsequent pandemic containment measures, and economic fallout have had a profound impact on people's employment choices, residential preferences, and travel habits. However, limited empirical studies have investigated longitudinal changes in the jobs-housing relationship, which is a critical intersection of these three elements during the global health crisis. Based on mobile signaling data from 2017 to 2022, this study employed visualization, employment activity compactness, and multinomial logistic regression to explore the multi-scale spatiotemporal evolution characteristics and causes of jobs-housing balance in Shenzhen during the COVID-19 pandemic. The results were as follows. (1) The spatial distribution patterns of residences and workplaces experienced the following evolutionary process: aggregation before the COVID-19 pandemic, dispersal during the strict control period of the pandemic, and, in the normalization phase, continuous dispersal in the first circle, as well as a rebound in the second and third circles. (2) Prior to the COVID-19 pandemic, jobs-housing relationships deteriorated; during the strict control period, employment self-containment remained stable, while cross-unit commuting distance significantly decreased, employment self-containment steadily improved, and cross-unit commuting distance returned to pre-pandemic levels. This trend is most prominently manifested at the 1 km grid scale across the three research scales. (3) Spatial heterogeneity affected the evolution trends of jobs-housing relationships before and during the pandemic, with public transportation accessibility and socioeconomic characteristics of residents being the main reasons for differentiation in evolutionary types. During the pandemic, regions with high dependency ratios, low average ages, and low educational levels exhibited delayed optimization of employment self-containment compared to regions with poor public transportation accessibility. Highly educated individuals were more likely to experience a continuous reduction in cross-unit commuting distances; migrant workers, older individuals, and female groups witnessed a decrease in cross-district commuting distances during the strict control period but a rebound extension during the normalization phase. The results contribute to a better understanding of the evolution of job-housing relationships during public health crises and provide support for future urban management.