Facing Covid-19 death toll curves

Pick your metric from the list and observe live total Covid-19 counts

When you just picked a metric or hovered a country, did you wonder about the actual meaning of what you saw? Is the scale of presented data making any sense to you? Did you consider it is about real people getting sick or worse, dying? Did you feel slightly uncomfortable with the previous question? How rational are you when you receive new information? How literate are you about data?

Data literacy covers the skills needed to extract knowledge from data.

We are living through the first world-wide pandemic with access to live data describing its evolution. Among the consequences, related news that reaches us have now two components.

The first is an old friend: fear. Fear of being sick, fear of seeing friends and family infected and so on. It is the natural emotional response that many among us are likely to feel.

The second is that you now know precisely how many people are hit, where and how: you are data-informed. If you do not clearly understand what the numbers mean, it is likely going to increase your level of anxiety. Most countries are using slightly different metrics. Is it about the case count? Or the percentage of ICU in use?

Media and governments combined are not likely to stop having confusing messages or playing on our fears. This means that navigating those messy waters falls upon everyone of us.

How then to boost your data literacy?

One way is to grow your critical thinking when you read numbers or charts. Do you ask yourself questions when you see a number? Does it have a scale? What is the reference or the average measurement? It is a rich topic, and I haven’t found a better reference on the subject than Han Rosling’s book: Factfulness. The book is full of insights to help you think with numbers distilled without (too much) maths. For a few years now, universities have communicated more broadly about these topics too. Calling bullsh*t from the University of Washington aims at teaching us constructive scepticism in a “data-driven world”.

Next, stop being passive with your sources of informations, select them. Going the extra mile will allow you to find sources that will be closer to the original source of information. They are much more likely to be factual and less emotionally loaded. Taking an online shopping analogy, did you ever click on a google ad to buy a product? Did it make you feel in control? The system suggest you how to behave. Looking by yourself for a product that is sold and manufactured in a manner that matches with your values is conversely a totally different experience.

The last point that I will share here is more about learning by doing. Nowadays, building visuals is getting really simple, whether you use online tools (e.g. infogram) or spreadsheet programs (Excel, Libre Office, etc.) or write down a few hundreds lines of codes like I did for the map above. The main point is not the technical part but to build it for someone, an audience. For example, creating the map above, here are some questions I asked myself:

• What do I want to show? (technical starter kit? what data people are curious about?, what would be useful for a more general audience?, etc.)
• What do I want to learn? (even if I do it in the spirit of sharing, what would I learn along the way?)
• Which source has the best quality of data? Are they accessible?
• How to best present the information (line charts, maps, etc.)?
• Is there an unbiased way to show the map? Where should be the middle of the map?
• What measurement will best represent the number of ongoing cases? Should I include a scale?
• Is it normal that country X has a zero count? (answers included: nobody lives there, the map file has a mistake in the name of the country, etc.)
• Is it normal that country Y has the largest death toll per million? (answers included: cross check different sources, follow the news, etc. )
• If I use such colour scale can everybody read it? Is it friendly to colourblinds?

Going through this process is educational for everyone, but you won’t be inclined to go through it if you do it in your corner. Asking yourself those questions becomes natural with experience and clearly separates emotional and rational responses.

Becoming data literate is a lot about: critical thinking, selecting your sources of information and building to share.

Media and authorities have to not only navigate a pandemic but also to face challenges about communication. How to convey knowledge in our data era? I have seen, for example, how the French government has changed its communication with the pandemic. It is not only about explaining and handling the emotional toll but also about building a story based on data and facts to justify the ongoing course of action. The style is clearly different and occasionally hectic. If anything, the Covid-19 pandemic can make us all better data literates and drives us to better informed decisions.

For the technical readers among you, below are shared all the code and data to reproduce the map.

Data sources

The data comes from:

• The country shape files have been downloaded on Natural Earth Data, exported to geojson format with Qgis and finally converted to topojson.
• Live Covid-19 cases count: disease.sh - Open Disease Data (aggregated from Worldometers).

Code

The code is provided in the following gist that you can access on github or directly on Bl.ocks (the data/support files are not shown below):