I'd have to disagree with Mr. Godin on this one. I think that getting the main point of Minard's graph - something like 'Napoleon's March on Moscow was disastrous' - happens very quickly, but like a good book, successive re-reads expose greater details that you weren't expecting. It does take time to read this chart in its entirety which is the joy for us data hounds but the main point is easily grasped at first glance.

"I don’t think that is what graphs are for. I think you are trying to make a point in two seconds for people who are two lazy to read the forty words underneath"

Seth Godin is out of his gourd and totally wrong. As Einstein said, you should make things as simple as possible, *but not simpler*.

I few days ago, you highlighted one of <a href="http://www.juiceanalytics.com/weblog/?p=254" rel="nofollow">Excel 2007's insane design decisions</a>, and seemed puzzled how they could justify what they did. You now have the answer: Microsoft, and their user base, are Seth Godins. I've always known them by the name of "pointy haired bosses".

Hallelujah, Seth!!!!! I think that more than some of Tufte's work is "the emperor's new clothes." The graph referenced in this email (Napoleon's March to Moscow) is a perfect example. I think we owe Tufte lots of credit for identifying the art of visualization as a real discipline or field that merits attention in its own right. But I think the field has moved on well beyond his work, particularly with new computing technology, and there's lots of terrific and amazing stuff that far surpasses anything he's done -- including the terrific work of Juice Analytics. (Full disclosure -- I have no association with Juice Analytics; in fact, I don't even know who these folks are; I was just lucky enough to have someone forward me their emails.)

Long live the iconoclasts!

I'm with Tufte (or at least, not with Godin) on this one. You do need a couple of words to explain what the main graphic is showing, but once you have that, the essential lunacy of Napolen's mission becomes very, very clear.

In order to satisfy Seth's requirement that a chart/infographic convey its meaning without any (or hardly any) contextual information, the chart has to either be so incredibly simple that it can hardly convey anything at all, or fall back on one of the well-known chart types (bar, pie etc) that everyone understands. If there's a vocabulary of visualization, that's like saying that everything has to be explained in words of one syllable.

The point of a visual is not necessarily to provide summary information but to clearly communicate an idea to the viewer what might be harder to convey via words.

Before making any infographic, you must know your audience. In the case of data analytics, I agree with Zach that the purpose is to quickly condense and provide summary information. In the case of Tufte's illustration, he is not presenting <strong>data</strong> but <strong>information</strong>. There's a big difference. Frankly, I rather spend 15 minutes absorbing his visual than reading black and white text.

Zach, I'm surprised you have not yet critiqued Dave Sifry's <a href="http://technorati.com/weblog/2006/11/161.html" rel="nofollow">State of the Blogosphere</a> graphs.

Godin' s two seconds vs. 15 minutes doesn't really allow for a serious discussion. IMHO, it doesn't matter how long you take to study a chart. The real question is how much useful data you put in it and how efficiently it communicates. If a high resolution, efficiently designed chart, shows you a general plot and invites you to study the details, probably it is a good chart, even if it takes you half an hour to study. Take a look at the diagram Death and Taxes: can you understand it in 2 seconds? How about any painting by Hieronymus Bosch (general concept, lots and lots of detail)?

I am not sure if Tufte is right. Today, we should be able to design something much better than this one by Minard. But we must fight this misconception that says that a) one must design charts for a 2-second attention span and b) if a chart isn't understood in 2 seconds that is a clear sign of bad design.

Simple graphs are made by simpletons for audiences of simpletons -- people who do not wish to trouble their pretty but simple minds analyzing complex data or complicated issues. The president's decision to go into Iraq provides the canonical example of this mindset. There are many simpletons in the world and not a few of them have lucked into positions of power, or aspire to do so. Hence, there is a great deal of demand for this approach and tools that facilitate its implementation.

Feh, Godin is part of the problem IMO, glitz with no substance. What's wrong with intelligent, thought provoking visuals. Yes, they may take longer than 2.5 seconds to digest but most things worth understanding do.

Godin's presentation is fun and amusing - the overall point being "Look at what you just did (built, designed, graphed) - does it make sense? Can it be better? Is it needed at all?" So, the friendly challenge to Seth and the Tufte critics is - show me your better graphic! If it is 'broken on purpose,' make it 'your job' to fix it.

Personally, I think the Minard graph is fabulous (and the best part of Seth's show is the soccer field sign...).

There's a significant difference between "graph" and "graphic" and I can't believe someone like Godin doesn't know this. Contrary to Godin's thinking, there are occasions when data can't just be dumped into Excel and graphed into simple little line graphs. Minard's graphic of a significant world event is one of those times. Is it complicated, yes; is it overly complicated, no. It's definitely not broken and if Godin doesn't like it, rather then just complain about it let's see a better "graphic" (or graph for that matter) that represents the data Minard published.

Talking about Minard's graphic in the same manner as talking about a road sign that only says "Caution: this sign has sharp edges, do not touch the edges of this sign" has got to be a joke. I can't fathom that Godin is that incredibly stupid.

There are different audiences. As Seth himself pointed out. He was being provocative, because he actually makes the case for Tufte later.

Sometimes, chart is like art. Inviting, charming, it gently invites your attention and slowly immerses you in it's study. Then there's a moment of liberation, a ha when you finally find out the multidimensional complexity behind a topic. A graph like the Napoleonic wars invites study better than a linear text.

The are other time when a chart is a picture-bite. A condensed, simplified point that hides the complexity. Not afraid to make it's point across. It's loud. It says here's the data. The world is black and white. But there's a time and a place for these charts too.

How about a mid-way compromise? I must admit I've used this graphic on my <a href="http://www.curved-vision.co.uk" rel="nofollow">presentation training skills courses</a> and got mixed responses. Some people think it's fantastic, something it's just awful. (Out of interest, I'm trying to look at their Myerrs-Briggs preferences 'cos I think it could be corellated.)

My current take is that it's got a lot of information and - once you've got your head around it - it's a great piece of work; but there's the rub as has been pointed out... you've got to get your head around it and work at it.

So surely the solution is to deconstruct it and show it incrementally in a series of building slides showing gradually more and more of the graphic? The solution to me would be to have a simple line showing the route but to have the wideline (the graph but) revealed a couple of months at a time, discussing the historical events that caused the width of the line (the number of men) to change.

Just a thought.

Simon