Quantifying human mobility behaviour changes during the COVID-19 outbreak in the United States

Abstract

Since the first case of the novel coronavirus disease (COVID-19) was confirmed in Wuhan, China, social distancing has been promoted worldwide, including in the United States, as a major community mitigation strategy. However, our understanding remains limited in how people would react to such control measures, as well as how people would resume their normal behaviours when those orders were relaxed. We utilize an integrated dataset of real-time mobile device location data involving 100 million devices in the contiguous United States (plus Alaska and Hawaii) from February 2, 2020 to May 30, 2020. Built upon the common human mobility metrics, we construct a Social Distancing Index (SDI) to evaluate people's mobility pattern changes along with the spread of COVID-19 at different geographic levels. We find that both government orders and local outbreak severity significantly contribute to the strength of social distancing. As people tend to practice less social distancing immediately after they observe a sign of local mitigation, we identify several states and counties with higher risks of continuous community transmission and a second outbreak. Our proposed index could help policymakers and researchers monitor people's real-time mobility behaviours, understand the influence of government orders, and evaluate the risk of local outbreaks.

Figure 1

Temporal changes of state-level Social Distancing Index. Figure aggregates the temporal change of SDI for the fifty states and the District of Columbia. The blue line shows the mean value of the state-level SDIs and the blue shadow shows the overall range. The grey dashed line marks the national emergency declaration in the U.S. The red triangular dots stand for the daily cumulative number of confirmed COVID-19 cases.

Figure 2

Social Distancing Index heatmap for all states. Figure shows the level of SDI scores for all states during the study period. Each pixel in the graph indicates the level of social distancing for one specific state on a specific day, where blue stands for more social distancing practiced and red for less. The “X” marker indicates the start date of state-wide, stay-at-home orders. The “O” marker indicates the order lifting date. The “I” marker indicates the start date of state-wide partial reopenings if different from the order lifting date.

Figure 3

Temporal changes of Social Distancing Index in the top five and bottom five states regarding the cumulative number of confirmed cases. Figure demonstrates the temporal changes of SDI scores in the top five and bottom five states in terms of the cumulative number of confirmed cases on May 30, 2020. The blue dots stand for SDI scores on weekdays and the orange dots for SDI scores on weekends. The red triangular dots stand for the daily cumulative number of confirmed COVID-19 cases. The grey line stands for the start date of the state stay-at-home order. The green line marks the stay-at-home order lifting date and the green dashed line marks the date of state partial reopening.

Figure 4

Temporal changes of Social Distancing Index in the top ten counties regarding the cumulative number of confirmed cases. Figure demonstrates the temporal changes of SDI scores in the top ten counties in terms of the cumulative number of confirmed cases on May 30, 2020. The blue dots stand for SDI scores on weekdays and the orange dots for SDI scores on weekends. The red triangular dots stand for the daily cumulative number of confirmed COVID-19 cases. The grey line marks the start date of state stay-at-home orders. The green line marks the stay-at-home order lifting date and the green dashed line marks the date of state partial reopening.