The limits of human mobility traces to predict the spread of COVID-19: A transfer entropy approach

Abstract

Mobile phone data have been widely used to model the spread of COVID-19, however, quantifying and comparing their predictive value across different settings is challenging. Their quality is affected by various factors and their relationship with epidemiological indicators varies over time. Here, we adopt a model-free approach based on transfer entropy to quantify the relationship between mobile phone-derived mobility metrics and COVID-19 cases and deaths in more than 200 European subnational regions. Using multiple data sources over a one-year period, we found that past knowledge of mobility does not systematically provide statistically significant information on COVID-19 spread. Our approach allows us to determine the best metric for predicting disease incidence in a particular location, at different spatial scales. Additionally, we identify geographic and demographic factors, such as users’ coverage and commuting patterns, that explain the (non)observed relationship between mobility and epidemic patterns. Our work provides epidemiologists and public health officials with a general – not limited to COVID-19 – framework to evaluate the usefulness of human mobility data in responding to epidemics.

Socioeconomic correlates of urban mobility trends in two Australian cities during transitional periods of the COVID-19 pandemic

“…Public health interventions put in place during the COVID-19 pandemic affected human mobility across all spatial scales, from the effects of international travel restrictions to the micro-distancing policies put in place to decrease crowd density in public spaces [2]. Due to the widespread use of mobile devices, changes in mobility that occurred during COVID-19 have been the subject of unprecedented levels of analysis and have proven useful for understanding the many factors influencing behaviour change during the pandemic [3][4][5][6][7][8][9][10][11].…”