Open-Source Platform for Population-level Surveillance of Personal Well-being and Mental Health
The unfolding of the COVID-19 outbreak was an unprecedented and unanticipated opportunity to understand how a sudden global shock modulates people’s online searches when seeking information about their emotional well-being. Furthermore, it also illustrated how public health surveillance systems were essential for tracking diseases’ spatial and temporal dynamics and shaping rapid public policy changes. The paper outlines a general framework to explore how digital epidemiology and machine learning can reveal aggregated human mental health and psychological distress expression measures. We also present an extract of results obtained in several current research exploring the relationship between big data time-series in the digital surveillance of search engines during the pandemic and a selection of social media feeds and official UK well-being surveys. The present body of evidence illustrates how data science can provide robust, finely grained, and replicable evidence on aggregated mental health measures at the population level. In the future, the digital surveillance method described here can be rapidly deployed to allow early detection of distress signals in a population to manage communication and policy action better.
The code contained in these two text files can be run in the R environment