Always a sucker for a good war and health related graphic, I can’t help but post about Nightingale’s Rose. The Rose is a graphical representation of William Farr’s mortality data on the Crimean War, drawn by the one and only Florence Nightingale. The chart visualizes infectious disease (in blue), battle casualties (in red), and other deaths (in black).
While interesting, the Nightingale Rose is, for three reasons, far from my favorite visualization of casualty data. First, the wedges are measured from the center (and therefore overlap), however this is not intuitively derived from the graphic itself and thus can easily lead to misinterpretation. Second, as with pie charts, I find it difficult to accurately compare wedges of similar size. Take, for example, the August 1984 1854 and November 1984 1854 blue wedges. Which is larger? Third, the area of each wedge represents its numerical value, however, since this is accomplished by increases the radius of the wedge, the representation takes on an exponential property. For instance, a wedge representing twice as many deaths as another wedge would not have twice the radius. Again, this takes away from the intuitive interpretation of the graphic. Not to say it isn’t still very cool.
This graphic is republished with permission from its creator, Hugh Small.
Note: This is an old post from 2008 that was reposted by accident. But, given that I enjoy visualizations, I am going to call this a fortunate accident.
Edit: Special thanks to Michael for pointing out that the Crimean War did not occur in the year of Reagan’s election.
Christopher Albon is a Ph.D. candidate specializing in armed conflict, public health, human security, and health diplomacy.
Want more? Subscribe to Conflict Health through RSS or email.
{ 4 comments… read them below or add one }
Your comments are valid, but you should also reflect the broader point of this graph: No comparing month to month, but comparing types of mortality to highlight the importance of medical care for troops, in particular for infectious diseases compared to battle casualties. The comparison should mainly be done within individual months, between types of deaths, not months to months. That was the aim, and it’s done splendidly.
Regards and thanks for the blog.
Did you mean August and November of 1854?
1984 might be a little late…
;)
Chris
I was thinking “gosh, didn’t he post this a couple of years ago . . . ?” I’ve been digging on these guys for a little while now, http://www.informationisbeautiful.net/ , kind of a ‘queer eye for the stats guy’.
Anyway, get to any world cup matches yet?
Florence Nightingale: The problem with intramonth comparison is that the wedge shape is not adequate for that purpose. Take March 1855 for example, it is impossible (or difficult) to tell just from eyeballing the graphic what the ratio of blue, red, and black.
Michael: Thanks for pointing that out, it is embarrassing to think how long that error has been there.
TEJ: It was my mistake, I was tinkering with the backend and republishing it. This visualization is beautiful, but not necessarily functional. Attended four World Cup games so far. Good times.