Using Mobility Data to Understand and Forecast COVID19 Dynamics

Abstract

Disease dynamics, human mobility, and public policies co-evolve during a pandemic such as COVID-19. Understanding dynamic human mobility changes and spatial interaction patterns are crucial for understanding and forecasting COVID-19 dynamics. We introduce a novel graph-based neural network(GNN) to incorporate global aggregated mobility flows for a better understanding of the impact of human mobility on COVID-19 dynamics as well as better forecasting of disease dynamics. We propose a recurrent message passing graph neural network that embeds spatio-temporal disease dynamics and human mobility dynamics for daily state-level new confirmed cases forecasting. This work represents one of the early papers on the use of GNNs to forecast COVID-19 incidence dynamics and our methods are competitive to existing methods. We show that the spatial and temporal dynamic mobility graph leveraged by the graph neural network enables better long-term forecasting performance compared to baselines.

Development of an Early Alert System for an Additional Wave of COVID-19 Cases Using a Recurrent Neural Network with Long Short-Term Memory

“…Few other studies have exploited mobility data, such as those generated by Google. For example, Wang et al [ 27 ] have shown that is of paramount importance to understand dynamic changes in human mobility, social networks and spatial interaction trends to better predict the still ongoing COVID-19 pandemic. Authors were able to demonstrate that incorporating Google-outputted mobility data resulted in a significantly higher predictive power of COVID-19 cases.…”

Deploying Vaccine Distribution Sites for Improved Accessibility and Equity to Support Pandemic Response

“…This question is often formulated as an appropriate variant of the facility location problem, which is wellstudied in the operations research literature (see Related Work). In our paper, we introduce a new variant that follows a recent line of work on integrating the mobility patterns of the population into disease models [Chang et al, 2021;Wang et al, 2020]. As is standard, we will use the distance from a vaccination center as the metric for defining accessibility.…”

Evaluating the utility of high-resolution proximity metrics in predicting the spread of COVID-19

“…Human mobility has been analyzed using multiple kinds of data, including cellphone pings [24], call detail records [15], social media geotags [47], aggregated mobility flows [42], and synthetic populations [40]. While previous research has focused on diverse topics, such as finding general patterns and laws of human mobility [18], analyzing traffic [4,31], or studying human movements during disasters [24,33], with COVID-19 and other epidemics, the emphasis has largely been on using mobility data to do epidemic forecasting.…”

“…Google released a public mobility index showing the change in mobility by region and activity type [16]. Google also shared aggregated mobility flow data with some research partners in a privacy-preserving way [42]. Similarly, Cuebiq released a mobility index and shared cellphone ping data with some research partners, also in a privacy-preserving way [11].…”

“…However, we take a more direct approach in the sense that our metrics are more of an intuitive indicator than time spent outside home in terms of how well people are maintaining social distancing. Wang et al [41] took a more involved approach by employing graph neural networks (GNN) to find the association between mobility and epidemic spread. However, GNN is computationally expensive and the question of explainability still remains.…”