Modeling the Spatiotemporal Association Between COVID‐19 Transmission and Population Mobility Using Geographically and Temporally Weighted Regression

Abstract

The ongoing Coronavirus Disease 2019 , which is caused by severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), has posed a serious threat to human public health and global economy (Yang, Sha, et al., 2020). Given the unavailability of specific drugs for the treatment of this disease, timely public health interventions (e.g., travel restrictions, social distancing and wearing of facial masks) are one of the most effective ways to prevent and control the epidemic. On January 23, 2020, Wuhan, the capital city of Hubei Province, China, implemented an unprecedented lockdown measure to prohibit people and motor vehicles from entering and leaving the city. Then, this policy was extended to the entire province of Hubei and other cities in China. Studies indicate that these control measures have greatly reduced the spread of COVID-19 in China. The Wuhan shutdown resulted in the delayed arrival of COVID-19 in other cities by 2.91 days (Tian et al., 2020). If the implementation of public health interventions was delayed for 5 days, the extent of the epidemic would have tripled in Mainland China (Yang, Zeng, et al., 2020).Population mobility is one of the major drivers of epidemic transmission, accelerating the dispersion of virus in space (Kraemer et al., 2019). Through the migration of population across regions, the virus is exported from one city to another. The association between population mobility and epidemic dynamics has attracted wide attention of researchers. Different data sources that record people's mobile behaviors, such as mobile phone, social media and public transport, can be used for this study (Yang, Sha, et al., 2020). Kraemer et al. (2020) quantitatively evaluated the impact of population movement on COVID-19 epidemic. Through linear regression modeling, they observed that the population movement data can well explain the distribution of COVID-19 cases in China. Jia et al. (2020) used mobile-phone data to track population flow from Wuhan and then correlated them with COVID-19 cases of 296 cities in Mainland China by using multiplicative exponential models. The results show that population flow data can be used not only to predict

Tracking and Controlling the Spatiotemporal Spread of SARS‐CoV‐2 Lineage B.1.1.7 in COVID‐19 Reopenings

“…(c) B.1.1.7 had significant spatiotemporal heterogeneity and human mobility played an important role in the transmission of B.1.1.7. This factor also supports the findings of some previous studies based on the relationship between mobility and COVID‐19 transmission (Chen et al., 2021; Jia et al., 2020; Levin et al., 2021; Li et al., 2021; Nouvellet et al., 2021; Sachak‐Patwa et al., 2021). The above conclusion can act as a guide to any region considering reopening to formulate precise entry measures and hence prevent the emergence of new variants in that particular region and thereby cutting possible networks of spread.…”

“…The results are shown in this study: (a) For all the 368 districts in Taiwan with high human mobility and low vaccine rates before the emergence of B.1.1.7, even the partial relaxation of the entry policy for specific imported groups, during the process of reopening, could lead to the emergence and rapid spatiotemporal 1.7. This factor also supports the findings of some previous studies based on the relationship between mobility and COVID-19 transmission (Chen et al, 2021;Jia et al, 2020;Levin et al, 2021;Li et al, 2021;Nouvellet et al, 2021;Sachak-Patwa et al, 2021). The above conclusion can act as a guide to any region considering reopening to formulate precise entry measures and hence prevent the emergence of new variants in that particular region and thereby cutting possible networks of spread.…”

Meteorological Influences on Spatiotemporal Variation of PM2.5 Concentrations in Atmospheric Pollution Transmission Channel Cities of the Beijing–Tianjin–Hebei Region, China

“…Additionally, the temporal and spatial variations of PM 2.5 concentration were strongly correlated with meteorological factors, and their relationships varied significantly across seasons and geographical locations, probably relating to PM 2.5 components. To better address this problem, GWR was applied in this study to capture the spatial nonstationary characteristics, but it could not simultaneously deal with temporal nonstationarity [62,63], which also could result in some uncertainty. Future research is required to elucidate these mechanisms.…”

A Novel Predictor for Micro-Scale COVID-19 Risk Modeling: An Empirical Study from a Spatiotemporal Perspective

“…On this basis, geographically and temporally weighted regression (GTWR), which takes spatiotemporal heterogeneity into account, was proposed [61][62][63]. Compared with the traditional GWR, which can only model in phases when processing data with both spatial and temporal dimensions, this often leads to biased and unsmooth results in time series, while the GTWR framework integrating the temporal autocorrelation of data is more advantageous in spatiotemporal epidemiological modeling [64,65]. It can be expressed as:…”