The NY Times COVID Spiral Is Off-Center
An opinion piece in the New York Times last week got a lot of attention in visualization circles for its use of a spiral chart as its opener. While the choice of chart and color scheme can be debated, I want to discuss the fact that the spiral is disconcertingly off-center.
The piece in question is titled Here’s When We Expect Omicron to Peak. The opener image, which attracted a lot of attention on Twitter, shows the number of new COVID cases in the U.S. since the beginning of 2020.
The initial response was largely negative, which then led to a wave of people defending it. I don't want to get into any arguments about whether liking it or not means you're narrow-minded, etc. – but I do want to point out what bothered me from the beginning, even though it took me a while to put my finger on it.
Spirals aren't uncommon to show periodic data, and the seasonal nature of COVID infections was clearly part of the inspiration here. We could argue whether the width of a ribbon is a good idea, and the choice of color was probably a large reason why this piece got such a strong reaction. But what bothers me is the wobbly spiral.
I've overlaid an Archimedean spiral on the chart below. It's centered on the same center as the one in the chart on the left, and shifted over to better match the piece on the right.
Now granted, there are many possible spirals, but the Archimedean one fits very well. But look at the one on the left, where I've aligned the centers: the one in the NY Times piece is shifted over to the left, which is what makes it feel really off-kilter to me.
When you look at where my overlaid spiral crosses the vertical axis, you will notice that its tangent (direction) at that point isn't horizontal. That might seem odd in a chart, where you're expecting the beginning of the year to lead cleanly into the next year. So my guess is that they decided to shift it over, which is what I did in the image on the right above. It lines up better, but not entirely. But now the whole spiral is shifted, which is what makes it look so wobbly to me.
Even when it's shifted, it looks like the lower right quadrant in both 2020 and 2021 is flattened relative to my overlay. That adds to the impression of the shift and wobbliness, I think.
I don't think the spiral was a bad choice here, but I do question the decision to shift it over like that. Perhaps using a different spiral would have worked better, or just accepting the beginning of the year to not be on a horizontal part of the spiral.
Posted by Robert Kosara on January 9, 2022.