Staying at Home Is a Privilege: Evidence from Fine-Grained Mobile Phone Location Data in the United States during the COVID-19 Pandemic

Abstract

The coronavirus disease 2019 (COVID-19) has exposed and, to some degree, exacerbated social inequity in the United States. This study reveals the correlation between demographic and socioeconomic variables and home-dwelling time records derived from large-scale mobile phone location tracking data at the U.S. census block group (CBG) level in the twelve most populated Metropolitan Statistical Areas (MSAs) and further investigates the contribution of these variables to the disparity in home-dwelling time that reflects the compliance with stay-at-home orders via machine learning approaches. We find statistically significant correlations between the increase in home-dwelling time (${\nabla }_{\mathit{HDT}}$) and variables that describe economic status in all MSAs, which is further confirmed by the optimized random forest models, because median household income and percentage of high income are the two most important variables in predicting ${\nabla }_{\mathit{HDT}}.$ The partial dependence between median household income and ${\nabla }_{\mathit{HDT}}$ reveals that the contribution of income to ${\nabla }_{\mathit{HDT}}$ is place dependent, nonlinear, and different given varying income intervals. Our study reveals the luxury nature of stay-at-home orders with which lower income groups cannot afford to comply. Such disparity in responses under stay-at-home orders reflects the long-standing social inequity issues in the United States, potentially causing unequal exposure to COVID-19 that disproportionately affects vulnerable populations. We must confront systemic social inequity issues and call for a high-priority assessment of the long-term impact of COVID-19 on geographically and socially disadvantaged groups.

2019 冠状病毒疾病（COVID-19）暴露并在一定程度上加剧了美国社会的不平等。在美国12个人口最多的大都市统计区（MSA），本研究大规模采集了人口普查街区组水平上的移动电话位置数据。通过机器学习方法，揭示了人口、社会经济变量与居家时间（HDT）记录之间的相关性，并进一步调查了这些变量对居家时间差异的贡献。HDT反映了对居家令的遵守情况。在所有MSA中，HDT的增加（∇HDT）与经济变量之间存在着统计上的显著相关性。优化的随机森林模型进一步证实了这一点：中等家庭收入和高收入百分比是预测∇HDT的两个最重要变量。中等家庭收入和∇HDT之间的部分依赖性，揭示了收入对∇HDT的贡献是随位置变化的、非线性的、在不同收入区间内是不同的。我们的研究显示，居家令是奢侈的，低收入群体则无法遵守居家令。居家令背后的不平等，反映了美国长期存在的社会不平等问题，有可能导致接触 COVID-19 的不平等和 COVID-19 对弱势群体的不成比例影响。我们必须面对系统性的社会不平等问题，呼吁优先评估 COVID-19 对地理的、社会的弱势群体的长期影响。

La enfermedad del coronavirus 2019 (COVID-19) ha puesto en evidencia la desigualdad social de los Estados Unidos, y, en cierto grado, la ha exacerbado. Este estudio revela la correlación entre variables demográficas y socioeconómicas y los registros del tiempo de residencia en casa, derivados de los datos de rastreo a gran escala de la localización de teléfonos móviles a nivel de grupos de manzanas censales americanos (CBG) en las doce Áreas Estadísticas Metropolitanas (MSAs) más populosas, al tiempo que adicionalmente investiga la contribución de estas variables a la discrepancia en el tiempo de residencia en casa que refleje el cumplimiento de las órdenes de permanecer en casa, a través de enfoques de aprendizaje automático computarizado. Descubrimos correlaciones estadísticamente significativas entre el incremento en el tiempo de residencia en la casa (∇HDT) y las variables que describen el estatus económico en todas las MSAs, lo cual también es confirmado con modelos de bosques aleatorios optimizados, debido a que el ingreso medio de la familia y el porcentaje de altos ingresos son las dos variables más importantes para predecir el ∇HDT: La dependencia parcial entre el ingreso medio de la familia y el ∇HDT revela que la contribución del ingreso al ∇HDT es dependiente del lugar, es no lineal y es diferente, dados los intervalos variantes del ingreso. Nuestro estudio revela la naturaleza ostentosa de las órdenes de permanecer en casa, las que los grupos de ingresos más bajos no pueden darse el lujo de cumplir. Tal desigualdad en las respuestas a las órdenes de permanecer en casa refleja cuestiones de vieja data en términos de desigualdad social en los Estados Unidos, que potencialmente estarían determinando una exposición desigual al COVID-19 afectando desproporcionalmente las poblaciones vulnerables. Tenemos que afrontar cuestiones de desigualdad social sistémica y pedir una evaluación de alta prioridad al impacto a largo plazo del COVID-19 sobre grupos geográfica y socialmente desfavorecidos.

Key Words:

• COVID-19
• mobile phone data
• random forest
• social inequity
• stay-at-home orders

关键词:

• COVID-19，
• 手机数据，
• 随机森林，
• 社会不公平，
• 居家令。

Palabras clave:

• bosques aleatorios
• COVID-19
• datos de teléfonos móviles
• desigualdad social
• órdenes de permanecer en casa

Acknowledgments

The authors thank SafeGraph for making this study possible by open-sourcing their mobility data sets. Thanks also go to the anonymous reviewers for their constructive comments and suggestions. Special thanks go to Meng Jin for her unwavering support.

Additional information

Notes on contributors

Xiao Huang

XIAO HUANG is an Assistant Professor in the Department of Geosciences at the University of Arkansas, Fayetteville, AR 72701. E-mail: [email protected]. His research areas include geospatial artificial intelligence, spatial modeling, deep learning, image analysis, and disasters.

Junyu Lu

JUNYU LU is an Assistant Professor in the School of Community Resources and Development at Arizona State University, Phoenix, AZ 85004. E-mail: [email protected]. His research areas include spatial statistics, climate change and adaptation, and extreme weather.

Song Gao

SONG GAO is an Assistant Professor of GIScience in the Department of Geography, University of Wisconsin, Madison, WI 53706. E-mail: [email protected]. His main research interests include place-based geographic information systems, geospatial data science, and human mobility.

Sicheng Wang

SICHENG WANG is a PhD student in planning and public policy at Rutgers University, New Brunswick, NJ, and an Instructor in GIScience in the Department of Geography at the University of South Carolina, Columbia, SC 29208. E-mail: [email protected]. His research areas include urban informatics, emerging technologies, mobility-as-a-service, autonomous mobilities, the gig economy, and transportation equity.

Zhewei Liu

ZHEWEI LIU is a PhD Candidate in the Department of Land Surveying and Geo-informatics at the Hong Kong Polytechnic University, Hong Kong, PR China. E-mail: [email protected]. His research interests include volunteered geographic information, big spatial data, and geospatial artificial intelligence.

Hanxue Wei

HANXUE WEI is a Doctoral Student in the Department of City and Regional Planning at Cornell University, Ithaca, NY 14850. E-mail: [email protected]. Her research areas include environmental quality of life, urban data analysis, housing market, and regional science.