QDS-COVID: A visual analytics system for interactive exploration of millions of COVID-19 healthcare records in Brazil

COVID-19 is responsible for the deaths of millions of people around the world. The scientific community has devoted its knowledge to finding ways that reduce the impact and understand the pandemic. In this work, the focus is on analyzing electronic health records for one of the largest public healthcare systems globally, the Brazilian public healthcare system called Sistema Único de Saúde (SUS). SUS collected more than 42 million flu records in a year of the pandemic and made this data publicly available. It is crucial, in this context, to apply analysis techniques that can lead to the optimization of the health care resources in SUS. We propose QDS-COVID, a visual analytics prototype for creating insights over SUS records. The prototype relies on a state-of-the-art datacube structure that supports slicing and dicing exploration of charts and Choropleth maps for all states and municipalities in Brazil. A set of analysis questions drives the development of the prototype and the construction of case studies that demonstrate the potential of the approach. The results include comparisons against other studies and feedback from a medical expert.