The proto-infovis specialist, Willard Brinton, has gotten his fair share of attention on the blogosphere over the last few years. With good reason too—this guy was way ahead of the game, writing about proper methods to be implemented while authoring maps, charts and graphs. Much of his writing, though, is pretty thinly veiled personal opinion. His map pin chapter in Graphic Methods for Presenting Facts (1914) reveals an extreme phobia of “the janitor’s feather duster.” So it should come as no surprise that he might have a few non-scientifically-based asides. Above is one such aside, a rant against the clock graph: This type of chart should be banished to the scrap heap. Wow. Such hatred for such a seemingly harmless graphic.
I’ll admit it; I first saw this graphic (and the comment about rectangular graphics being easier to make and read) months ago. And it’s been on my mind since. I just couldn’t figure out why Brinton would have so much distain for it. To be honest – as long as the variation between observations is not minute (in which case, why would there be a graphic to begin with?) – I actually quite like the look of this type of graphic. It’s decently easy identify a seasonal trend, is it not? But then I ran across Charts and Graphs (1925), by Karl Karsten. And look at what he has to say about these graphs.
As a chart, this is worthless. Great. So this really piqued my interest. What is the deal with these clock graphs!? And why don’t I hate them when I really, really… should? Well, it turns out some are more successful than others. Check out some examples (all pre-1970).
Arthur Lockwood, in Diagrams: a visual survey of graphs, maps, charts and diagrams for the graphic designer, refers to these two images as “diagrams”, the first showing temperatures throughout the year and the second showing “employment of working hours on a weekly basis for a group of sample farms in Finland”. Successful or not, it’s pretty clear that Finnish farmers are really into the earth’s crust (though I’m not sure that’s the point). Is it this kind of muddled graphic that fueled Brinton and Karsten’s hatred? I look at this lower image and think I may have been caught up in the aesthetic novelty of the others if I was actually getting ready to defend the clock graph… because, I’m sorry, this Finnish Farmer graph is unnecessarily difficult to read.
From Calvin Schmid’s Handbook of Graphic Presentation (1954), here’s another clock graph. I think this one mostly works… but now I’m starting to agree with Brinton. Does the polar coordinate system really add anything to this data set? And as for Karsten’s hatred… well, if you root around a bit more in his book, you’ll find the following two clock graphs.
I think the bottom graph is fairly well readable. But the top one gets a little muddy. Does it have too many subsets of data? It might not, if there is any meaningful correlation. But is there? To me, it simply looks like this department store sells more stuff on Saturdays in December. Do we need a graph for that? And if we do, do we need it to be a clock?
I still think there is a place for these charts. I really do. But after looking closer at those made during the age of Brinton and Karsten, I completely understand the distain.
Read Full Post »