Evaluating the Effectiveness of Digital Social Mobility Data in COVID-19 Predictive Models: A Scoping Review (Preprint)

BACKGROUND The COVID-19 pandemic has been devastating the world since late 2019. In response, governments across the world implemented mobility restrictions to halt the spread; however, these limitations only slowed transmission. To understand the disease and guide public health interventions, there have been many attempts to develop predictive models. One method to forecast COVID-19 is through using Machine Learning (ML) and Artificial Intelligence (AI). Therefore, this paper seeks to review the literature for articles using mobility data as an indicator for ML-based COVID-19 forecasting models. OBJECTIVE This paper seeks to review the literature for articles using mobility data as an indicator for ML-based COVID-19 forecasting models. METHODS Four databases (Scopus, Embase, Web of Science, and PubMed) were searched for articles related to forecasting COVID-19 using ML and mobility data. The articles were then screened and excluded based on a series of inclusion and exclusion criteria. Once the articles were screened, aspects related to the data used, ML models used, and outcomes were extracted from the article. RESULTS This study examined and identified four patterns of findings. The first was that most of the studies included examined the USA. This paper also found that Long-Short Term Memory (LSTM) was the most common and effective model when forecasting COVID-19. The most common mobility dataset used was Google, however, SafeGraph and Baidu’s mobility datasets also produced quality results. CONCLUSIONS From this review, it appears that mobility data is a good indicator for ML-based on COVID-19 forecasting models. Despite this, there are a lot of improvements that can be made such as the inclusion of COVID-19 variant data. It is also evident in this review that the most common limitation in this field is lack of data, such as lack of public health compliance data, and the asymptomatic nature of COVID-19