My friend Jimmy of Code for Excel and Outlook twittered about a chart on TechCrunch that showed the general online assessment of the Olympic coverage by NBC. How We Hate NBC’s Olympics Coverage: A Statistical Breakdown shows an analysis of nearly 20,000 tweets and 5,700 blog posts. The highlight of this analysis is the following chart:
Actually, this isn’t the original chart, theirs was substantially larger. Since theirs was obviously constructed in Excel 2007, I transcribed their data and built my own. Though smaller, this chart lacks none of the features of the original.
Jimmy made some remark in his tweet about the “worst chart ever”, but I have to say, this chart is not even close to the worst ever. It may be in the bottom quartile, but we’re talking about a long, long tail.
So what’s wrong with the chart, anyway?
We’ve heard many times about how people just aren’t very good at judging angles or areas, and that makes pies ineffective for all but the simplest parts-of-a-whole displays. Donuts take this one step further, cutting out the central bit of the pie, so we’re relying on areas alone, without any help from the angles where all the wedges meet.
Of course, the donut resembles the big fat zero most viewers would give NBC as their grade.
Let’s improve this chart in steps. First, if we sort the data points, we only have to try comparing adjacent points. This also puts the most biting criticism of a sports event, “Not Enough Sports”, right up front.
Now we can put the munchkin back into the donut. We have the angles to help us judge the areas, which may or may not help.
Finally, we can convert the pie pieces into candy bars.
Now the labels are right next to the data points, not off in some distant legend, and the bars are easily ranked by length. The above chart may cause confusion with its multicolored bars, and we don’t want any viewers hunting around for the key to find meaning in the bar colors where none exists, so we use a single color for the bars. Or in this case, two colors. I’ve highlighted “Happily Watching” in a distinct color to set this category off from all the negative ones.
The above chart has two sets of labels. There is the horizontal axis at the top of the chart and the labels at the end of each bar. Is it redundant to have both in the same chart?
We can choose to leave data labels off the points and rely solely on the axis labels.
Or we can keep the data labels on the points, remove the axis, and close up the space between the chart and the title.
Which labeling option to you prefer? Axis labels, data labels, or both?
UPDATE 3 March 2010
Steve Fleming suggested in a comment below that I move all of the data labels between the category labels and the category axis. Good idea.
Similarly, the labels can be moved from the ends to the bases of the bars.
Jeff Weir says
Couple of thoughts:
– I’m not a fan of the white datalabels. Too hard on my failing eyesight. (Too much chart porn, I guess)
– If you get rid of the numerical axis, you can get rid of the grid lines too.
– To get completely ‘presentation zen’-ish about it (a great book by Garr Renolds if anyone’s wondering) you could get rid of the horizontal axis line too.
Here’s how that would look:
Out of curiosity, anyone notice anything different about this graph?
Mike Woodhouse says
I like the final, data label, version. For one thing, it saves me from having to interpolate by eye. At least, should I actually care about the difference between 12% and 13% when I got my Olympic coverage from the BBC.
I’d agree that the main point to be taken from the data is the 85%-15% negative-positive split. While the bar chart is clearly superior to the ghastly pie/donut, it’s still not (to my eye) showing the proportions as expicitly as it might.
I came up with this (XL 2002)
Jon Peltier says
Jeff – In the first chart you had a subliminal border on the bars (a slightly darker shade than the fill), which I sometimes use to make the bars look crisp. The second chart lost the borders.
Mike – In this case, the stacked bars also works well, given the alternative interpretation of like/dislike. Who cares why it was so bad, it just was.
Thom Mitchell says
Axis labels, data labels, or both?
I prefer both. As Mike wrote, the data labels save me from a visual interpolation. At the same time, so few people seem to know about effective and honest chart design that I would like to see the axis labels so I know that the chart designer isn’t playing games by *not* starting the bars at zero.
Mike’s stacked bars give me the additional information that all the information is being presented (because the percentages sum to 100), but in this instance I would be more interested in comparing the “Too much” items and the “Too little” items against each other. So the bars have it!
Annie Pettit says
Just happened across your blog and i’m quite enjoying it! Glad to see people working hard to eradicate chart misuse. :)
JP says
I knew it wasn’t the worst chart ever, but it got your attention. I thought the donut style was distracting, for the reasons you specified. I’d prefer to see the “Happily Watching” group integrated with the rest in the sort order, but I understand why NBC would want it singled out.
Candace says
Jon-
Love your suggestions, and really think you made some great improvements. Thanks for helping us, and am hoping you will see some of these suggestions integrated into our next release. Personally, I prefer your last version with the white data labels. We’ve been talking about our data display for awhile, and you nailed it.
Thanks!
Candace
CEO Crimson Hexagon
Steve Fleming says
Another style would be to move all of the data labels into a column to the right of the list of complaints and to the left of the bars. That would facilitate scanning between the labels.
derek says
For me the answer depends on how many data points there are in the chart, and whether the quantities are displayed as positional or retinal variables, to use Bertin’s terminology. For bar charts with few data, I will consider abandoning the scale and just naming the points. For line charts with a zero base and few data, I will consider changing to a bar chart. And for line charts with a non-zero base (which therefore can’t be a bar chart) I don’t consider abandoning the scale.
I notice your health care chart yesterday had eleven labels on the scale, when in my opinion six or even four would have been more than enough. I think the smallest possible number of tick labels is three: one at the bottom, one at the top, and one in the middle to show it’s not logarithmic, or to show that it is. At work, I usually have four or five; the most is six, because they’re part of a set, and because I’m also forced to have a dual axis. I compensate by being generous with unlabeled ticks, though I understand some disapprove of lots of ticks–strangely sometimes the same people who lard their scales with numbers!
I wish Excel had better support for three-level tick schemes (labeled major, unlabeled major, and minor), and I sometimes roll my own scale numbers just to get that look. I once had a chart with a horizontal scale of a century, where each year was exactly discernible, with only six labels: five intervals of twenty years between labels, ten years between unlabeled major ticks, and two years between minor ticks. Any point lay either on a tick, or between ticks.
Jon Peltier says
Candace – Welcome to my blog. I’m so glad you took my comments in the spirit they were intended (unlike a previous inspiration for one of my posts).
Steve – Good idea. See the update to this post.
Derek – I’ve been accused of overdecorating my axes. It’s more a case of not overruling the defaults, but it’s unnecessary ink anyway. Another problem is having percent or dollar signs on all labels instead of just the top one, which adds to the reader’s eyestrain. Again, it requires overriding the default formats.
Adam Goldsmith says
I got hung up in Jeff Weir’s “Olymipc” typo, but I think that’s more a reflection on me. Personally, I think Jeff’s graph works best in a presentation, but either Jon’s final version or Mike’s stacked graph would work great in a report.
Jeff Weir says
Just those pesky grid-lines to go now, Jon:
Or too minimalist?
Matt Healy says
My biggest beef with their coverage was not among the listed categories: what annoyed me the most was that unless at least one US or Canadian citizen was among the medal contenders the event was unlikely to get televised, and unless at least one person from the US or Canada was on the podium the medal ceremony certainly wasn’t gonna get televised. Are there any readers familiar with coverage in other places, who can tell us whether their networks showed similar biases?
Jon: the complex scoring system used for figure skating would seem a cool place for you to try some data mining…
Jon Peltier says
Matt –
I saw an infographic last week comparing the scores of the top six make figure skaters, with respect to the new scoring rules. It was almost interesting for a hockey fan. I tried to Google but couldn’t find a link.
Andrew says
Interesting article.
This confirms what I’ve already worked out, when I go abroad (from the UK) the television coverage is generally dire.
Good ol’ BBC!!!
Americans make the best spreadsheet software though, even if the default charts are naff…