Some days ago, Martin Elmer (@maphugger) tweeted Numberphile’s video in which they highlight the famous Minard map. Martin and I exchanged a series of tweets on the topic:
Martin had some substantial criticism of Minard’s graphic. Maybe, graphic is the crucial word here: A part of Martin’s and my discussion was about the (too?) subtle geographic contextualisation the map provides. While Minard labels his work as “(une) carte” (French for map), Edward Tufte‘s famous quote on the topic says that Minard’s depiction “may well be the best statistical graphic ever drawn” (emphasis mine). Telling? Or not?
In any case, Martin has followed up on our discussion and provided some links to interesting discussions around, and resources about, Minard’s map (Thanks!). I will list these here and encourage you to check them out:
- Minard was an engineer. “Virtually all of his maps were published privately, and made their way into the public domain haphazardly.”: Charles Joseph Minard: Mapping Napoleon’s March, 1861
- (Martin:) “Before he depicted Napoleon’s army, he made Hannibal’s: with a much more map-like depiction.”
- Revisions of Minard’s map.
- “Additional options to tell the story (…) rely[ing] on the sophistication of PowerPoint as a production tool” (I think the author is not being sarcastic): Gene Zelazny responds to Tufte on the famous Minard graphic. I don’t like the proposed options.
- Cruel Pies: The Inhumanity of Technical Illustrations, by Sam Dragga and Dan Voss, contains this memorably paragraph about Minard: “
Nowhere in this visual display is
the slightest indication that the subject being illustrated represents the slow dying of 412,000 human beings. (…). We don’t see the people, and we don’t see their anguish.”
The human dimension
Now, especially Dragga and Voss (in the last bullet point) raise an interesting concern with Minard’s and similar depictions: the elegant abstraction that opens the information to relatively easy appreciation also removes, in the opinion of the authors, (too much of) the human angle to the story, namely the tragic loss of lives.
The authors suggest two potential remedies: use of pictographs and overlay of statistical graphs on photos of human subjects. The reworked Minard map in Figure 7 does not work well for me (bad alignment of, and not ideal choice of type and scale of pictographs) – but: I can relate to and do sympathise with the need for ethics in data visualization. As unusual and ornamental (in a bad sense) some of the proposed graphic additions may seem, they may, at the same time, augment memorability (but that’s opening a whole new can of worms).
Text [from ‘texere’: to weave]
Throughout the discussion between Martin and me I kept thinking about the text that goes with Minard’s map. To me (German-speaking Swiss, but still, Swiss: thus, you know, expected to speak some French, too), consuming the text (at least the title) came quite naturally as part of looking at the depiction. The two, map and text, seemed interwoven. To Martin this was different and we agreed that it might have caused part of the differences in our opinions.
Abstracting from the discussions about the actual language used in such annotations, what can the merit of such text be? Can it too help to build the ethical dimension of visualizations? Many will probably argue that an image is much more powerful than words, but still, I am not ready to give up the idea that through accompanying powerful text we can bring humanity into visualizations.
Given his thoughtful treatment of language and cartography in Maps and Prose, I’d wish Martin would tackle this topic at some point.
What next?
Finally, and coming back to the Minard map, it turns out that there is an open MSc project with Arzu Çöltekin at the University of Zurich Geovisualization and Analysis lab centred on both Minard’s map of Napoleon’s march and John Snow’s Cholera map. If you’re as curious about the (overestimated? rightly lauded?) effectiveness of these visualizations as me, why not make use of this unique opportunity to scrutinise these hallmarks of geographical visualization using latest technologies such as eye-tracking.
In any case, I’m looking forward to further research into Minard’s and similarly widely distributed and cherished visualizations.
I’m flattered to see my spate of tweets turned into a blog post! The whole topic is definitely one that deserves more than 140 characters, so I’ll take this comment as an opportunity to articulate my thoughts a little better.
The first thing that must be acknowledged: as far as we know, Minard’s map wasn’t commissioned by anybody, or formally published in any capacity, we can’t really say who its audience was, or what its purpose was supposed to be, or what it intended to communicate. That makes it practically impossible to judge whether the map is a success or a failure. But it’s always interesting as a discussion piece.
Regarding its use of text: I support the use of text when it used to clarify and contextualize the content within the map (or whatever the information graphic is). If, however, the text exists just to explain how to read a map that would be unintelligible otherwise, it’s probably a sign that the map needs a more intuitive graphical representation for its content. Minard can be forgiven on this point: pictoral legends didn’t really exist in 1869, so we can probably assume that a 19th century reader would benefit more from a textual explanation more than a modern-day reader.
Another thing that I’ve realized regarding the language barrier: the English word “map” is rather ambiguous: it is also used to refer to other sorts of diagrams and charts that aren’t “maps” as Cartographers know it. And the phrase “figurative map” is essentially meaningless in English. I suspect that the French “carte” is less ambiguous in this way, and the title would better cue Francophone readers about the graphic’s abstracted geography.
My second major criticism is one I never really expanded upon in my tweets, even though it was the entire reason I made that first comment in response to the Numberphile video. Minard’s graphic is heavy on *data*, but short on *information*. Tufte praises the map for representing six different quantitative variables, but… so what? What do I actually learn from the combination of those 6 datasets that couldn’t be conveyed via some other, more conventional method? Zelazny made some junky PowerPoint charts, but his executive summary is spot on (albeit tersely worded): “THE COLDER IT GOT, THE MORE SOLDIERS DIED”. A picture is supposed to be worth 1000 words, but the map of Napoleon’s march only manages those 8.
This goes hand-in-hand with the criticism of the map’s inhumanity. If this were truly “the world’s greatest information graphic”, it would convey to the reader much more of the things they would need to or want to know about the campaign. This includes an understanding of the campaign’s human cost. It might be my personal bias, but I am much more impressed by information graphics that holistically educate their readers about the chosen topic, rather than just dropping a bunch of data onto an image.