http://blogcritics.org/ A sinister cabal of superior bloggers on music, books, film, popular culture, politics, and technology - updated continuously. en Copyright 2005 by the authors Mon, 1 Aug 2005 11:25:13 EDT http://backend.userland.com/rss Blogcritics.org custom software http://blogcritics.org/archives/2005/07/29/080001.php#comment-194303 I'm a fan of Tufte. He tends to focus on much more complex charts; I feel that the media can do a better job even with the simpler charts. But I'm happy that NYT and Economist are among the few publications that actually believe in the power of visual graphics. The response will be brought over soon. 194303@blogcritics.org Mon, 1 Aug 2005 11:25:13 EDT http://blogcritics.org/archives/2005/07/29/080001.php#comment-194161 Kaiser, with your interest in representation of numbers, I recommend the books of Edward Tufte to you - check them out. Also, if the response is sufficiently detailed, bring it to blogcritics. 194161@blogcritics.org Mon, 1 Aug 2005 02:01:04 EDT http://blogcritics.org/archives/2005/07/29/080001.php#comment-194157 Phillip, I just posted a response on my blog. You made me think more about this. 194157@blogcritics.org Mon, 1 Aug 2005 01:51:48 EDT http://blogcritics.org/archives/2005/07/29/080001.php#comment-192526 Kaiser, I think that the idea of a pie (or donut) chart instead of a chart that uses line segments is to compare to dissimilar quantities on an equal basis. That is, they work best to represent percentages. Consider: a line chart such as the one you created, above, has different lengths of line overall. In fact, glancing at your chart, it would be easy to think that the population is expected to shrink by 2030, but that isn't the case. In fact, the population will grow, but the lines say otherwise. In addition, the segments are cumulative, which seems confusing. With a pie/donut chart, I can tell at a glance the relative sizes of the segments more easily than I can with the line chart. My mind tends to think that the Hispanic segment of the popular is expected to grow from 2$ to 6% based on the size of the gaps between the tick marks. In fact, what would happen if two segment were to cross? It just so happen that in this chart all segments are growing except the top and bottom, and the top is far enough in the lead that the shrinkage doesn't cause it to encounter the next highest segment. But if, say, the Hispanic population were projected to shrink from 14% to 10%, your chart would descend into chaos. In fact, the only way that I can use the proferred alternative chart is to look at the numbers themselves to make sense of it, which somewhat defeats the purpose of using the chart at all. The non-textual, non-numeric visual cues run completely counter to the actual data, so why have them at all? There are undoubtedly ways to improve the original chart. Spinning them around so that the words can appear between paired segments rather than being connected by big lines might be one. And there may be ways to salvage your chart. But the bottom line is that percentages are best represented with a chart style that is uniformly sized, and you don't get any more uniform than a circle. They all have exactly 360 degrees. Fascinating topic, though! Thanks for the post. 192526@blogcritics.org Fri, 29 Jul 2005 14:24:59 EDT http://blogcritics.org/archives/2005/07/29/080001.php#comment-192337 fascinating topic - thanks and welcome Kaiser! 192337@blogcritics.org Fri, 29 Jul 2005 09:36:02 EDT