In the end of November 2021, German health authorities registered the 100,000th dead person as a consequence of the ongoing Covid-19 pandemic. On the occasion of this sad landmark, we created a map of Germany showing every death by the location where it has been registered. But not all deaths are shown at once. Instead, viewers are encouraged to scroll through time from March 2020 until winter 2021 while seeing more and more deaths burning as little dots into their screens. Eventually, to complete the rather emotional setting, we added short stories underpinned with fitting photos.
In a short amount of time, the article reached an above-average number of readers and provoked numerous comments directly on the website. Some commentators where literally scared by the sheer number of corona deaths appearing on their screens. Obviously, the dot-density representation of the data coupled to the user-controlled time scale opened a new perspective to many readers. Usually, our readership is more comfortable with the standard death-rate curve that does not show individual deaths. One reader got quite affected by the scroll mechanism: “My fingers are sore from scrolling.” Thus, by letting the users control the addition of deaths, we succeeded in fueling the convenient data with emotions.
Next to the scrollytelling experience we created an animated video to be shared via Twitter and Instagram. On the latter platform, for example, we reached more than one million views.
After all, we faced the challenge that a large majority of our readership knows corona fatality data inside out. The curve of daily deaths is continuously exhibited on the spiegel.de frontpage and presented by many other news outlets as well. This and the rather impersonal nature of a statistical curve gives rise to psychic numbing. Here, we tried to bring back the human stories behind the numbers and the public reactions on our website, in letters to the editor and multiple social platforms seem to prove us right.
We got daily death numbers by county from the Robert Koch Institute, our in-house data journalism team and – for some early death cases – through own research. The data was cleaned and aggregated using R with RStudio. The frontend of the article was built with Svelte. Selected D3 functions were embedded in Svelte components to for example draw the map outline and the line chart. The Turf.js library allowed us to randomly position dots within a county. The 100,000 dots were drawn on canvas through WebGL using the Regl library to allow for the best possible performance.
What was the hardest part of this project?
Taking a publicly well-known data set and shedding a new light on it can be challenging. The idea of a map of Germany with 100,000 dots was born quite quickly. But soon we got some doubts and questioned ourselves: Will our readers understand how the visualization works? Will there be misinterpretations? Or are they maybe just ignoring the new perspective?
In a nutshell: We needed to make sure that we always focus on the raw numbers and underpin them with built in and well-defined emotions.
Certainly, the hardest part of the project.
Our solution was to couple the time axis with the mouse wheel and touch fingers. Then we personalized the Covid-19 deaths by using one dot per person. And, with the map, we added a geospatial personalization layer. Thus, people could really feel how the death rates evolved. It is real people behind each dot. They lived somewhere in Germany. Even close to the hometown of our readers.
Surprisingly, the subsequent addition of dots by scrolling solved an issue many dot-density maps come with. They are often just population-density maps in the end. However, here, we were able to show the death rate dynamics over time rendering additional layers of information on a simple dot-density map.
What can others learn from this project?
This project taught us to think outside the box. Dominated by the rapid news cycle we produce veritable amounts of graphics and visualizations all day. But sometimes we forget about our readers. We hope that this project – published on one of the most visited web pages in Germany – encourages other data visualization designers in the newsroom to come up with creative and user-centric ideas as well.
After nearly two years in this pandemic, we urgently need to develop new ways to access the vast amount of data that we collect every day. Standard charts and curves are important to make policy decisions, but they are far away from the social reality of our readers.