« Disproportionate Risk | Main | Swans and Geese »

Robbins and Few

robbins.gif Creating More Effective Graphs
Naomi Robbins
Wiley-Interscience, 2005 • ISBN: 047127402X

This is a perfectly sensible, reasonable book, but it irritated me. First, it’s not that original: there’s extensive quoting from Cleveland’s books, including reproduction of many examples, while Playfair and Minard make their usual appearances, second-hand from Tufte. Recycling of previous authors’ examples seems to be endemic in this literature; Tufte even recycles himself. Perhaps recycling examples fights global warming. Anyway, Robbins’ debt to Cleveland is notable in the use of dot charts, and the proliferation of boxes around elements. I wouldn’t say the figures are particularly innovative or well-designed, and most of the text is just common sense, like “large markers are easier to read than small ones”, and “overlapping things are harder to read” (I’m paraphrasing, or course—the style is not this pithy.) There’s actually surprisingly little content; often a short paragraph will float in the middle of a page, padded with ornamental margin art, and a page or two after the graph it’s discussing. The whole thing could have easily been a quarter of the length. Plus it’s over US$40 and contains no color, which, given the fundamental importance and frequent misuse of color in amateur graph-making, seems unforgivable. So not a recommended purchase, particularly to the readers of this blog who may well have come across all this advice already.

few.gif Information Dashboard Design
Stephen Few
O’Reilly, 2006 • ISBN 0596100167

Few has a seemingly quite specific aim: producing a series of small charts and tables that fit onto a single screen—an information dashboard. It’s aimed at, but not limited to, businesses who need to summarize performance indicators at a glance—many of the principles are quite general, and in fact scientists presenting data would do well to think in terms of a cluster of small figures that present related information, rather than a few big independent ones. There is plenty of good sensible advice on toning down garish colors, reducing chartjunk, trimming away superfluous labels and gridlines and so forth, in much the same vein as Tufte. I wouldn’t say it’s an essential book, particularly if you already have a work or two by the usual suspects. Nor does it add much that’s new. It does, though, recontextualize good practices, differing from Tufte in that the writing is very clear and straightforward, and so far more likely to be tolerated by a down-to-earth managerial or administrative type. The best strategy would be to show them the chapter on simplifying charts, then Few’s elegant high-density redesigns at the end, so they leave you alone and stop asking for colorful banners and animated logos.

Comments

You are spot on with the Robbins book. Really light on the content and just does snippets from other works. Not a ton of originality. If you have the Cleveland books, this is not necessary.

I like Few's works. He starts with a little physiological background into human perception, and the approach is well grounded, unlike Tufte's more esoteric writings. Few doesn't care what you use to make your charts, even *gasp!* Excel, while Tufte implies that if you don't use highfalutin' academic packages, you're wasting your time.

Comments Welcomed

Verification (needed to reduce all the spam I get):