3 ms·
Even the people who are supposedly masters at this type of data presentation mess it up big time: http://blog.jgc.org/2009/11/how-to-fail-at-data-visualization.
by jgrahamc 15y ago
Even the people who are supposedly masters at this type of data presentation mess it up big time: http://blog.jgc.org/2009/11/how-to-fail-at-data-visualization.html http://blog.jgc.org/2009/11/how-to-fail-at-data-visualizatio...
- chimeracoder 15y ago> He uses circles with dots in the middle to represent the size. So is the value being plotted proportional to the radius of these circles or the area? Not clear It's worse than that. Is the size of the circle proportional to the number of incidents, or to the density of incidents? Large circles look like they should the most 'dangerous', but that would imply a high density, which would imply a small circle. The entire visual representation serves no purpose except to contradict the data it presents.
- DanBC 15y agoHe spent a year creating over 200 visualisations. So, 4.5 graphics a week, for a 46 week year, makes 207 graphics. I'm not convinced that people can gather suitable data, analyse it, turn it into an attractive graphic, and then check the graphic for statistical or numerical or logical or whatnot errors.
- tel 15y agoI find there are roughly two clusters of infographics people. One side is the statisticians who are adept at handling large data and making reasonable inference from it. Their primary goal is to tease out a story from the data. Unfortunately, they often don't care if it's ugly or impenetrable so long as it's "correct" or at least strongly defensible. The other side are the designers. Their primary goal is to explore the attractive nature of patterns in data and use those patterns to tell people an interesting story. Unfortunately, if the data doesn't seem to fit the story they want to tell then they're likely to fall prey to confirmation bias and other errors of detail that can completely change or pervert the support for their story. Neither side is successful in isolation.
- vosper 15y agoMy former employer produced a lot of the second type of infographics; despite having an analytics department the focus was solely on cute visuals, the actual data was invariably just a percentage breakdown of some group, and they repeatedly broke basic rules of charting. But the most annoying thing, by far, was that not one of them conveyed information more intuitively or more quickly than a list of numbers or a standard Excel chart would have.
- essayist 15y agoSomewhat parallel to what I see as the two legitimate uses of infographics (with no claim that there are only two): 1. presenting/supporting analytical results 2. providing the viewer/reader with a way to explore data for new conclusions (cf Tufte) The first one is easier. E.g. I've identified that there were successive waves of support, in the Republican primaries, for Bachmann, Cain, Gingrich, etc. and I can show a chart drawn from the Nate Silver do support this - http://fivethirtyeight.blogs.nytimes.com/2011/12/25/reads-reactions-16/ http://fivethirtyeight.blogs.nytimes.com/2011/12/25/reads-re... The chart both shows and supports the conclusion. Another "type 1" infographic in a poster/narrative style (via Tufte himself): http://www.edwardtufte.com/bboard/q-and-a-fetch-msg?msg_id=0002w4 http://www.edwardtufte.com/bboard/q-and-a-fetch-msg?msg_id=0... The second use is more tricky. The famous "Napoleon's campaign against Russia" chart does it well - http://upload.wikimedia.org/wikipedia/commons/2/29/Minard.png http://upload.wikimedia.org/wikipedia/commons/2/29/Minard.pn... You can explore lots of different points. (And it works in the first sense as well - hitting you between the eyes with how many French soldiers died). My test for the first purpose is "what's the story?" If it's just "ooh, look at the pretty", the infographic has failed. And then of course you verify that the story is presented fairly. My test for the second purpose is whether the infographic rewards more examination, and, even more so, whether I can find patterns or links that the author/graphic artist might now have anticipated. E.g. in the Minard/Napoleon graphic, you could check how losses en route to Moscow correlate with losses on the way back. Given how often infographics fail both these tests, I'd say we're in an infographic bubble. Can't wait for it to burst.
- tel 15y agoI agree with all you've said here. Infographics is fashionable today in the same way that every hour there's a new article on how Big Data Will Change Everything. As an article of fashion, it's easy to see that people have a real desire to understand their world. So the trick is just to prove that you're actually better at it than everyone else. Tufte has done this, I think. In some circles.