Time-series clustering for home dwell time during COVID-19: what can we learn from it?

Abstract

In this study, we investigate the potential driving factors that lead to the disparity in the time-series of home dwell time, aiming to provide fundamental knowledge that benefits policy-making for better mitigation strategies of future pandemics. Taking Metro Atlanta as a study case, we perform a trend-driven analysis by conducting Kmeans time-series clustering using fine-grained home dwell time records from SafeGraph, and further assess the statistical significance of sixteen demographic/socioeconomic variables from five major categories. We find that demographic/socioeconomic variables can explain the disparity in home dwell time in response to the stay-at-home order, which potentially leads to disparate exposures to the risk from the COVID-19. The results further suggest that socially disadvantaged groups are less likely to follow the order to stay at home, pointing out the extensive gaps in the effectiveness of social distancing measures exist between socially disadvantaged groups and others. Our study reveals that the long-standing inequity issue in the U.S. stands in the way of the effective implementation of social distancing measures. Policymakers need to carefully evaluate the inevitable trade-off among different groups, making sure the outcomes of their policies reflect interests of the socially disadvantaged groups.

Analysis of communities of countries with similar dynamics of the COVID-19 pandemic evolution

“…In this regard, as Ehlert (2020) points out, policymakers should consider the identification of risk groups when designing their response strategies. In addition, Huang et al (2020) recommend policy makers to guarantee that the impacts of public policy include the interests of the socially disadvantaged groups. It is also important to consider the existence of trade-offs within the objective function, since minimizing both COVID-19 propagation and economic impacts might be two conflicting goals.…”

Cell phone mobility data and manifold learning: Insights into population behavior during the COVID-19 pandemic

“…Our results are consistent with other studies linking demographic characteristics to cellphone mobility data during the 2020 SARS-CoV-2 pandemic. Two recent studies using data from SafeGraph found mobile devices from areas with higher median household incomes stayed home more than devices from lower-income areas [9, 11], and this trend occurs in other mobile device datasets [7]. These studies hypothesize that the relationship between income and mobility is due in part to the ability of people with high-paying jobs to work from home.…”

“…Several related studies cluster mobility time series by a single demographic characteristic selected a priori , such as income [7, 9, 11] or population density [10] or party affiliation [8], to demonstrate behavioral differences with respect to that characteristic. Alternatively, one could reduce the time series to a summary statistic, such as average stay-at-home level over a particular time window, and study the relationship between that metric and several demographic covariates.…”

“…The approach has been integral to revealing the heterogeneity, but also the consistency, of mobility patterns across California, Georgia, Texas, and California; it has enabled a multi-scale geographic perspective on behavior allowing insights at the state, urban-rural, peri-urban, and suburban scale. Recent efforts have also utilized clustering of mobility time-series data specifically for analyzing SafeGraph stay-at-home data in Atlanta [9]. Our approach, though, is substantially broader in scope; identifying the low-dimensional embedding of the data enables a characterization of the geometric structure and the relatedness of each CBG mobility behavior.…”

“…Mobility data offers a unique measurement instrument to link public health statements and related legislative actions taken to reduce population mobility with an actual effect on population behavior. Cell-phone mobility data has provided early evidence that these orders were indeed associated with reductions in movement [7, 8, 9, 10, 11]; moreover, adherence was not uniform and may be associated with factors such as socioeconomic status and political leanings [7, 8, 9, 10, 11]. The keen interest in cell-phone mobility data to help inform policy makers during the COVID-19 pandemic has been widely discussed [12], with strong emphasis on the challenges facing data ascertainment bias, interpreting the link between mobility and behavior changes, and the lack of a single mathematical framework for analyzing this data [5, 6].…”