What is wrong with these charts?

July 15, 2013  |  Ugly Charts

Speaking the world

Whoa. There are a lot of things wrong with this chart. Gold star for every mistake that you find. And there are many stars to hand out.

While we're at it, check out this record-breaking house cat. The owner must spend a fortune in kibbles and litter.

Housecat

The reasoning? The bar chart's value axis starts at 13 feet six inches instead of zero, which is a perfect reason of why you're not supposed to do that. Now, it would be a different story if you changed the chart to show clearance between the top of an average trailer and bridge. The cat comparison would make sense. [Thanks, Kurt]

45 Comments

  • Apparently more than a trillion people speak Chinese. Which is unusual since that exceeds the human population by a pretty good factor. And I don’t speak Chinese. You would think I would have to for that to even be close to correct… hmmm.

    • I think the comma is a decimal point, but still, switching units doesn’t make sense.

      Also I think the bar’s height and width is supposed to have some meaning, but it looks like the area ratios of the bars are not correct.

      • Good point, in many European languages, “1,026 billion” would mean one point zero two six billion. …Actually that just means it’s a second-language mistake and proofreading mistake, rather than a typo and a proofreading mistake… thinking about it, that’s probably the same level.

        As for how the bar lengths thing happened… no, I’m stumped. I thought maybe it could have come about by the columns of bars getting out of step with the column of labels, but no: nudging either up or down or adding or removing a missing row won’t fix it.

        The cat thing is funny… Though they could have made it work if they’d made it look like the cat was sitting on a clearance bar and made it clear that the chart’s bars continued below towards zero, with a squiggle to suggest that it’s been cropped – then it would actually be quite useful.

      • I’m positive that that the comma should have been a decimal point, but the point (pun intended) is that precision is important. If a comma were to be replaced with a decimal point in my paycheck, I would be rather outraged.

      • oneblankspace August 7, 2013 at 7:23 am

        In many of those languages where a comma is used for the decimal instead of a point, a billion is a million million rather than a thousand million like in the US.

    • I gave them benefit of the doubt on the billion – the comma is a decimal separator in some locales.

      The area of the speech bubbles seems to be all fubar, regardless.

    • Plus… we’d have to invent a language called “Chinese” maybe it’s a burp… then we’d be closer. I guess.

    • Besides that, none of the bars appear to be drawn using the same scale, regardless of what arbitrary base may be desired.

  • Matthew Vose July 15, 2013 at 9:21 am

    466 million is only marginally less than 765, but a lot more than 380, according to the bars.

    Not sure what the depth of colour is supposed to be about either :S

  • 1,026 billion speakers? That would make the bar much longer compared to 765 million speakers below…even assuming they meant 1,026 million speakers, I don’t think those bars are accurate lengths.

  • The bars are odd as well. See English and Spanish – the bars are almost identical but the values differ with ~300mil.

  • Bar lengths are certainly not matching the figures, unless a very weird horizontal scale is used! The top bar suggests he might be using area instead of length as a measure, but it still doesn’t match the figure. And why using different colors for the bars? Maybe that’s the key of his secret encoding!!!

  • Andrew Burton July 15, 2013 at 9:34 am

    Whoa. Awful, awful.

    Let’s start. What do the speech bubbles represent? Nothing. Even worse, the boxes in the speech bubbles look like they should represent something quantitative (width or area), but they don’t. The “English” speech bubble box (SBB) is maybe a tiny fraction bigger than the “Spanish” SBB, but there are (from the numbers) about 65% more English speakers than Spanish speakers.

    The Chinese SBB is more than twice as tall as the others, no reason why. And there aren’t 1,026 billion speakers of Chinese – there are 1.026 billion, or 1,026 million. BTW, no need to repeat “million speakers” – could be noted at the bottom.

    The chart is supposed to be accurate to within 4 significant figures – but it’s undated, with no source. We can’t tell if the designer used multiple sources, or a single source, and we don’t know (if multiple sources were used) whether they’re the same date. We also don’t know if we’re talking primary or primary plus secondary language. My guess is primary.

    The weird giant figures are also non quantitative. And, FWIW, I’m not a fan of mixing font sizes for the same set of information (looks like there are three font sizes for the language names, with “Portuguese” and “Indonesian” having different sizes to the other names.)

    That’s all I got.

  • The author did not measure the bar length well – also, the Chinese language bar has double the thickness of the others?. The graph looks like sticks of swiss cheese.

    • “The graph looks like sticks of swiss cheese.”
      Definitively the worst part, I totally agree ;)

  • 1) same as aaron
    2) why is the bar for chinese speakers so thick?
    3) what are the sources?

  • Alex Reutter July 15, 2013 at 9:57 am

    The semitrailer/bridge graphic should perhaps instead show the clearance (i.e., difference between bridge height and max. semitrailer height, so that the first bar is 11 inches, the second bar is 7 inches, etc).

    Then again, since the purpose is to show how insanely little clearance there is, perhaps the scale should start at 0 and include a graphic of a 13’6″ semitrailer trying to go under each of the bridges. With the cat on top of the trailer, of course.

  • To be fair, they were probably experimenting with using a logistic scale for the bar chart between “Bengali” and “Arabic”, decided they didn’t really like the way it looked, and then realized that “Hey, that tree scale thing doesn’t look so bad” between “Spanish” and “Chinese”.

  • I see a few problems with the chart. We can’t assume any of the figures are correct as there is no reference. The bars are way out of scale compared to the numbers. The speech bubbles are meaningless and therefore confusing. Billions are compared with millions, which is also confusing. The colour progression of the bars has no relationship with the numbers. The graph gives the impression that nobody is multilingual or that there might be many dialects within a language. All of the illustrated figures are white males of very strange proportions.

  • Bars aren’t to scale. 123 million speak Japanese. The difference between English & Spanish is 300 million, but the yet the difference between the Eng & Span bars is much smaller than the Japanese bar.

  • Are you looking for “the shades look like the urine chart the Boy Scouts give out, so the Chinese must be really dehydrated”?

  • Nathan,

    Have you seen the original at
    http://10-most.com/the-top-10-most-common-languages/

    It’s still ugly, but the bars match the numbers. It looks as though someone edited the labels without changing the bars.

    • @Thomas. Thanks, the original chart seems to make more sense.

    • The order of languages in the original chart is different.

      • @Eoms: Yes, they’ve edited the language names and the numbers, but not the units (as you can see by the font) or the bars.

        Also, the numbers on Nathan’s version look like ‘total number of speakers’ and on the original version look like ‘number of native speakers’, based on Wikipedia, so the edit is plausible.

    • Amelia Bellamy-Royds July 23, 2013 at 12:35 pm

      And suddenly it makes sense…

      As you say, still ugly, still has the questionable grouping of all Chinese languages and the unspecified data source, but at least the bars match the numbers. It could have been more clear if the three-times-as-thick bar for “Chinese” was broken into three separate bars, but that’s my only critique on the data visualization front (everything else is just garish design).

      For the monster kitty, in contrast, I think a little extra artistic design could have helped communicate the author’s idea. If the shaded box for the “maximum semi-trailer height” was actually made to look like the top of a truck, with the cat sitting on it, and maybe the cat could be wearing ski goggles and have its ears plastered back on its head as it hurtles towards a black-and-yellow maximum clearance sign representing the “ideal clearance height”, then I think I might understand it at a glance.

  • A couple that jump out at me (aside from over a trillion people speaking Chinese which has already been mentioned)

    * The chart makes English and Spanish look very close. (Like Bengali and Russian). The numbers say otherwise
    * The taller bar for Chinese. Which seems to imply they’re meaning for the areas of bars to represent the differences in the numbers, rather than just the length.

  • What Aaron said, plus, y’know all Chinese speakers are Asian. And “Chinese” is one language. (Your surname would be Yiu in my dialect ;) )

    The bars are not to scale, and the colours confuse me. Ug-lay.

    Where are the sources for these numbers?!

    P.S. The average cat size to scale is great though. Monster kitteh!

  • The guy in blue, has an eye on his forehead.

  • Any chance the poster’s author is European (in which case we’d need to read the comma as a decimal point, as is the norm in the EU)? Hard to give him or her the benefit of the doubt, though, given the other howling errors. My personal favorite has got to be the visual implication that 765m is approximately equal to 466m, but 466m is massively greater than 380m.

  • Disturbingly unproportional with changing bar widths. Bar lengths don’t match up, no wonder there is no x-axis since it would be hard to show a difference of 300 million as tiny while a difference of 100 million is huge. Colors change for no reason and the graphic is oddly racist.

    Do I earn a signed copy of Data Points? haha

  • The scale of the bars are off from Bengali to Arabic (assuming bottom scale correct), and then Spanish to English. Plus what is the ‘double’ height bar for Chinese?

  • Well spotted by Thomas there! I was wondering about the squashed-ness of the 1026 as an aside and that totally explains it.

    Reminded me of this: http://en.wikipedia.org/wiki/Hypergraph versus http://www.hyperedge.com.au/why-hyperedge

  • Arun Viswanathan July 15, 2013 at 10:38 am

    Three mistakes I could spot:
    (a) The units are inconsistent (billions for Chinese).
    (b) The size of the bars seems very arbitrary (for example, the bar for Spanish speakers should roughly be slightly less than half in size compared to the Chinese).
    (c) It is unclear why the bars are colored in different shades of gold.

  • 1. The language names are wrong
    2. The numbers are wrong
    3. The numbers lack citation
    4. The units are wrong
    5. The font sucks
    6. The colors suck
    7. The bar area is not proportional to the numbers
    8. The speech bubles convey no useful purpose or data
    9. Why is there a period after “world languages”?
    10. The text/bar verical spacing is random
    11. The numbers don’t add up
    12. The people are all men (don’t women speak?)
    13. The men are not proportional to any data
    14. The men have 3D shadows, which is impressive for 2D people
    15. Fonts appear to be randomly sarif and san sarif

    This makes most amatures look professional by comparison. BLECH!

  • The only other thing I would add to Harvey’s list is that the “million speakers” is needlessly duplicated and adds to clutter. I’m actually sorry I ever saw this chart!

  • Um, how about the fact that Chinese isn’t a spoken language. Mandarin and Cantonese are the most common “dialects” of Chinese, but really they are both separate languages unto themselves. Also, there is no clear definition for “Most Spoken Language”… are they counting people who speak English or Spanish or Hindi as a second language? Also, the people’s heads are way too large for their bodies.

  • Jasin Blaze July 15, 2013 at 5:55 pm

    Not every Chinese person speaks Beijing dialect. I’m pretty sure the numbers reflect native speakers not speakers as many people in the world speak more than one language.

    • Most mainlanders speak the Beijing “dialect,” but it may not be the majority of Chinese citizens first language (although that is changing as Mandarin is the first language of Chinese youth). The problem with calling it a dialect is that the “dialects” are mutually unintelligible with each other – which makes them separate languages. http://bit.ly/276ThS

      You are totally right that these only reflect native speakers. I knew that many people spoke a second language. What I didn’t realize was that there are more multilingual speakers than monolingual ones! http://bit.ly/85m5AB

  • A better title might be “What are the 10 most WIDELY spoken languages in the world?”
    We have no idea on whether the Chinese are more talkative than the English….

  • Gurupad Hegde July 18, 2013 at 7:09 pm

    “Multilingual speakers outnumber monolingual speakers in the world’s population” Source:wikipedia. Nothing is mentioned about this overlap.

    Nothing mentioned about age group considered.

  • Horzontal and vertical scales do not make sense, nor do the color progression or the “bubble” fillers of the bars. No source for the data. Number scale is not consistent (billions, millions?). The giant people cartoons add more “noise” to an already crowded design.But the worst offense is, not explaining what they mean by “speaker” and not differencing between Mandarin, Cantonese, etc. (these are alll under “Chinese”). When they say “speaker”, do they mean “native speaker” or are they including second-language speakers? Because there are far more Spanish native speakers than native English speakers, but if you count ESL, then you wind up with that number of English speakers.

  • (a) Lengths of the bars are not linearly proportional to the number of speakers
    (b) Chinese spoken by millions and not billions
    (c) Aspect ratio of the Chinese bar changes

Copyright © 2007-2014 FlowingData. All rights reserved. Hosted by Linode.
7ads6x98y