Since we launched the this Coronavirus dashboard a few weeks ago (March 16, 2020) we've had an enormous response, most of it very positive and supportive. Naturally there are some people who have questions about the data and the dashboard itself, and there are those that think we should be doing more. So we put together a small list of frequently asked questions to explain some of the processes behind the creation of the dashboard. We'll add to this as things progress.
Who developed and maintains the dashboard?
The dashboard was originally developed by the Media Hack Collective, a small data analysis and visualisation team based in Johannesburg. Media Hack released the first iteration of the dashboard on March 16, 2020, and then shortly afterwards partnered with the Bhekisisa Centre for Health Journalism. Bhekisisa is doing some of the best reporting on the Covid-19 epidemic in South Africa which is why we partnered with them. Bhekisisa is assisting Media Hack with the collection of data (see below for the challenges related to this) and we publish links to their stories in our newsletter.
How many people are working on the dashboard?
Media Hack is a team of 2 people and we're collecting and sorting data and building the various visualisations on the dashboard.
How often is the data updated? Why is it so slow to be updated?
We get this question a lot. There are a number of answers to this. The dashboard is typically updated once a day. The Department of Health (DOH) usually issues a statement once a day and this is our primary source of data. These statements are not released at regular times so sometimes they are released in the morning or, often, late at night. We update the dashboard as soon as we have new information. The DOH statements and data made public by the National Institute for Communicable Diseases (NICD) cover the information on the dashbord that is specifically local, such as the number of confirmed cases, deaths and the provincial breakdowns. The data for the charts that look at a more global perspective (eg. those covering Africa and the 100 days charts) are sourced from both local and international sources and are often more detailed than the local data. They include, Johns Hopkins CSSE, European Centre for Disease Prevention and Contol, Worldometer and Our World in Data.
Why don't you have gender/age breakdowns on the dashboard, or more detailed data on local municipality infections?
The short answer is that we simply don't have that level of data. It is not provided by the Department of Health or the NICD. Unfortunately there is no live stream of detailed data that is publicly available (or at least not that we're aware of). There are internal projects to government that may well have more detailed data but we don't have access to that data source. When the first confirmed cases of coronavirus were announced the Department of Health released very detailed statistics including local versus international sources of infection, as well as gender and age breakdowns of those confirmed infected. As the number of cases increased the data became less and less detailed.
The Western Cape government has released some lower level data but for us to add it to the dashboard it needs to be consistently updated and released. The City of Johannesburg has also released some more detailed data, but not consistently.
We have no relationship with the Department of Health so while we would love to see more detailed data being made available we don't have any agreements with data providers that would give us access to more detailed data. We rely on Bhekisisa to provide us with official media statements as soon as they receive them.
Who pays you for producing the dashboard?
Short answer: no-one. We are producing the dashboard as a free public resource. We have other jobs and work that we do to earn a living but producing the dashboard is not one of them.
Can we use your data? Do you have an API?
There are two answers to this: (1) you're welcome to use the data we have on the dashboard but (2) we aren't able to provide a data feed (such as an API). The data we're using is manually collected and stored in about a dozen different spreadsheets, and these are changing or being added to all the time depending on what charts we're showing at the time. While we're being very careful about how we collect and use the data, the process around this is very fluid and we don't have the resources to manage any kind of data feed service.
However, a lot of the data we do use is publicly available thorugh a number of global organisations, including: Johns Hopkins CSSE, European Centre for Disease Prevention and Contol, Worldometer and Our World in Data. And for more localised data the Data Science fo Social Impact Research Group at the University of Pretoria has been collecting data on Github
Listen to the podcast
The nice folk over at Volume featured us in the latest episode of their Media Diaries podcast. Each episode looks at independent media organisations working through the Coronavirus lockdown in South Africa. In the episode we talk about the inspiration for the dashboard, how we did it, and some of the challenges keeping the dashboard up to date and useful. Take a listen:
Any other questions?
For more information on the work we do at Media Hack please visit our website: mediahack.co.za
To read Bhekisisa's fantastic coronavirus news coverage visit their website: bhekisisa.org