A research group based at the University of Pretoria (UP) is playing its part in disseminating factual information about the novel coronavirus that the world is currently in the grip of by setting up a databank that offers users real-time updates about COVID-19.

The Data Science for Social Impact group aims to use data science as a way to find solutions to social problems and to get end-users and decision-makers to better understand how to use and understand data science and its limitations. For the current health crisis, it has set up a repository and dashboard that pulls information from the National Institute of Communicable Diseases (NICD) and the Department of Health (DoH).

The manner in which the information is presented will hopefully empower South Africans with accurate data and statistics, says Dr Vukosi Marivate.

The data is presented in an easy-to-interact format on a dedicated site. ABSA Chair of Data Science at UP Dr Vukosi Marivate leads the group and says the databank was set up to pool resources in a time of crisis.

“Very early on we saw that the coming epidemic was going to stretch the country’s data capabilities,” says Dr Marivate who is a senior lecturer in UP’s Department of Computer Science. “Once the minister of health and the NICD started publishing their data only in statements, we thought about how other researchers may need to get hold of this information in a more accessible way. As such, group members and collaborators have worked to build tools to automate the data gathering and cleaning as much as possible. Validation is also done through discussions about errors and rectifying them as soon as possible.”

The group also set up the repository and database in order to provide an interactive dashboard that the public could use to better understand the impact of the pandemic on the country.

The research group consists of more than 15 members, including masters and doctoral students, other research organisations as well as affiliate members who regularly collaborate with UP.

The databank was set up to pool resources in a time of crisis.

“One member of the research group, Herkulaas Combrink, has a health background and advised us on some of the needs of health professionals,” adds Dr Marivate. “As it stands, the project has become very volunteer-driven, with many volunteers from South Africa and beyond lending their skills. We now have maps and an API [application programming interface] courtesy of these volunteers.”

The group is, however, cautious about making predictions about the progression of COVID-19. “We believe that predicting the spread of the virus is best left to epidemiologists at the moment,” says Dr Marivate. “We look to the leadership of NICD and the South African Centre for Epidemiological Modelling and Analysis for their experience in disease modelling. We need to be responsible. Many predictions are being made, but we need our energies to get us through this. We do not want to add to any unnecessary panic or get sidetracked. We would like to have the best data to be available to experts.”

Another challenge of the COVID-19 pandemic is curbing the spread of fake news in relation to the disease. Dr Marivate says the transparency of the data research group and the manner in which the information is presented will hopefully empower South Africans with accurate data and statistics.

“A big part of our design and discussion with regards to the dashboard was on it being simple to navigate and being factual,” he says. “As such, we update only from NICD or DoH reports, so anyone can check our raw data to confirm how we come up with our visualisations and numbers. We are very transparent in our process.”

The data repository can be accessed at https://github.com/dsfsi/covid19za and the dashboard at https://bitly.com/covid19za-dash.