Based on my teaching experience, I know that many people will find this Euler diagram very confusing. I've been thinking about how to teach interpretation of these kinds of diagrams for a bit. So, for folks who might be confused, here's a thread on how to read this... https://twitter.com/EricTopol/status/1250420486086881280
With these diagrams, you're interested in how different ellipses overlap with each other (or don't overlap). For ex, section labeled "259" is overlap of dark orange (loss of smell), light orange (cough), and grey (fatigue). Note that dark & light blue and black don't overlap.
This indicates that 259 people reported experiencing loss of smell, cough, and fatigue, but did not report diarrhea, shortness of breath, or fever.
If you want to know how many people reported loss of smell, you need to add up all the numbers in dark orange ellipse. 140 + 259 + 69 + 26 + 17 + 18 + 23 + 41 + 50 + 64 + 17 + 46 + 281 = 1,051 There were 1,702 respondents, so that's why header says ~60% report loss of smell.
These types of diagrams are used a lot to compare gene content among a set of genomes. I've published papers with Venn diagrams comparing gene content among genomes...I think they can be useful, but here, there are so many overlaps that the diagram is very busy and hard to read.
Fortunately, someone has already posted a different way of looking at these data using a plot called UpSet. https://twitter.com/ngehlenborg/status/1250307072861720583
In this plot, you can see each symptom listed at the bottom, with a bar showing how many people reported that symptom. See the dark orange bar for loss of smell? That shows slightly over 1,000 people, which is what we calculated from the Euler diagram.
The neat thing about UpSet plots is how they show intersections. Remember that we noted 259 people reported loss of smell, cough, and fatigue. This corresponds to the second column in the UpSet plot. Note how the dots are filled in for each of those symptoms on the bottom panel.
The UpSet plot also shows more clearly than the Euler diagram that fatigue, loss of smell, and cough are prominent symptoms. These symptoms on their own or in some combination account for the first 5 columns on the UpSet plot.
I think this is a great illustration of the advantages of UpSet plots over Euler diagrams. Here they are side-by-side. My lab works on bacterial genomics, and I'm planning to use UpSet plots in the future to help us explore genome similarities and differences.
You can follow @MicroWavesSci.
Tip: mention @twtextapp on a Twitter thread with the keyword “unroll” to get a link to it.

Latest Threads Unrolled: