Like deciding who won the Indy 500 by picking the car with the snazziest paint job

Daniel Lakeland writes:

My wife sent me this link, saying how cool it looked. I [Lakeland] told her it was one of the worst things I’d seen in a long time…Apparently it won the Guardian’s “Visualization Contest”…

The map *appears* to plot number of plane crashes per country, totally missing the point of expressing these as a rate (ie. crashes per 100k flights or probability of crashing or something). I pointed out that there might be countries where 100% of all 3 planes flying into that tiny country crashed yet it would look like the best place to fly into…

And then they report “ratio of fatal accidents to number of planes in services (the higher the better [sic])”. Better if you’re a terrorist perhaps….

The list of atrocities goes on…

Another visualization from this same site has a more subtle problem:

This one compares things that don’t have the same units. Some of these are dollars per year, some are total dollars ever spent on some thing, some of these are dollars of capitalization per company… Let’s not even get into the issue of how were things inflated or deflated into current dollars, and who did the estimation of some of these numbers…

Just goes to show, statisticians and other quantitative folks should make our visualizations a lot more beautiful, because we don’t want people to look at extremely poorly thought out but beautiful visualizations and say “wow I saw this great data about how terrible the US airline industry is and how much we spend on all these different things…” when they could be looking at really well thought out data analyses and say “wow I saw this beautiful graph of this really interesting well thought out analysis of global spending…”

I [Lakeland] applaud this guy for his enthusiasm for visual design. I hope he takes a course in data analysis before he published more books on visualization.

This reminds me of a blog discussion I had with Nathan Yau. Pretty pictures are nice and even thought-provoking, even if they mangle the data. I prefer the George Orwell approach, but others like it more flashy. So, even though I agree 100% with Lakeland’s comments, I’m wary of criticizing those visualizations. For the goal of conveying information, they’re horrible, but for sparking interest in their topics and motivating readers to look carefully at the numbers, maybe they’re ok.

To give them prizes in a visualization contest, though . . . well, I wouldn’t do that. That would be like deciding who won the Indy 500 by picking the car with the snazziest paint job.

I’ll have something more to say about this soon.

1 thought on “Like deciding who won the Indy 500 by picking the car with the snazziest paint job

  1. On the "Reducing your odds of dying in a plane crash" page, they have a chart for the actual odds of dying in a plane compared to other ways of dying. But they don't compare flying to anything other form of transportation. To be fair, they should have the death odds for cars, bikes, and walking. Instead the compare the odds to entirely different categories.

    That page is wrong in so many ways. . .

Comments are closed.