Very pretty graph missing a y-axis
As this graph loads into the browser frame (since, for some reason, the static data needs an animation to appear onscreen), note that there’s something missing: a label for the y-axis.
There’s one clue to what the graph is measuring, but it’s available only if you physically roll the mouse over one of the data points. Only then do we see the word “adoption” - oh, the graph is talking about user adoption of these technologies. But by whom - Americans, or North Americans, or all humans alive at that moment - is left to the imagination of the reader.
Thus we see that cell phones are just a hair under 52% (why is the division on 52% and not 50%? no guess)… but whether that means 52% of all Americans have a cell phone, or 52% of everyone in the world has a cell phone, we have no way of finding out.
I remember the good old days when y-axes were labelled, and when graph were stored in quick-loading graphics. Get off my lawn!
Would a bar graph have been so difficult?
Try making sense of this jumble of numbers. Taken from the NYT magazine on Sep 30, 2012:
WSJ and a sloppy Apple marketshare graph
From the WSJ, Apple Tightens Up on Apps contains this chart of Apple’s marketshare.
From a quick glance, it appears to show a steady decrease in marketshare, from 100% down to a fraction of that.
Only after concentrated study can we see that the bar graph isn’t telling a story at all; instead, it is sloppily displaying four data points. There’s no need for the bar graph (except perhaps to fill the page with some visual color).
How should one display four data points in two categories? With two lines of text, as such:
- Smartphone: 20% US, 14% international
- Computer: 12% US, 5% global
A sloppy Michael Bay analysis
In an otherwise entertaining analysis of Michael Bay’s movies comes this graph, showing the cost and gross of all his films.
To see what’s wrong, try comparing Pearl Harbor to Bad Boys II. It’s not immediately clear that they cost almost exactly the same amount, yet Pearl Harbor grossed a bunch more. The problem lies in the near-impossibility of quickly comparing the area of two rectangles. It’s much easier to compare length… in a BAR GRAPH.
The visual design is so pretty (such nice right-justified margins!) that it compromises what readers can actually learn from the graph. Sloppy.
A bar graph would have been so much clearer. What we have here is a failure to communicate. (I’m speaking of the data, not the films, though you could certainly make a case.)
P.S. The work overall is strong, as some of the other graphs on the page are actually quite expressive - in particular the final graph, which shows a correlation between number of explosions and total gross.
Microsoft’s vision of the future: more sloppy info
Microsoft just launched a video showing their vision of the future - full of touch screens and mobile software in every conceivable context. If anything like this future comes to pass, we’ll have plenty to write about here!
Consider the vision of the credit card-sized mobile device. The interface sports an apples-and-oranges assemblage of “hubs,” news headlines, project updates, and a “health tracker.” Look in the lower-left to see the amazing futuristic info-visualization UI tracking your health:
Admittedly this is just a speculative vision showing rough ideas for future interfaces. Microsoft doesn’t mean this literally. But the Health Tracker is a good example of the UI thinking throughout the video: tons and tons and tons of data, presented in varying shades of blue and gray. Lots of smooth futuristic feel, with very little design toward understanding.
Along the same lines is the mockup of a cupcake recipe. No need for readable type or colors with any contrast - just smooth, gray-on-gray outlines of ingredients. How much sugar, how much flour? See if you can tell:
Is this really our future - more and more data, presented in sleek fonts and delicate color schemes? That’s sloppy. Microsoft should think about how to build for better understanding, not just how to display more (sloppy) info.
P.S. More discussion on the video over at Metafilter.
WSJ goes nuts with MTA data
In this WSJ piece we see that “a year’s worth of data shows how people move around the city, and how a 7.5% fare increase changed habits.” The article then includes an interactive infographic showing the data.
It’s way too much data - hundreds of subway stops, each with its own bar graphs - with almost no analysis or presentation of conclusions. Could someone explain what we’re supposed to learn?
For example, take a look at the two graphics below: one for 116th Street, and the other for Avenue J. Both have bar graphs showing similar shapes - not many senior discounts, how interesting - while a call-out reminds us of one datum from Flushing.
While it’s an impressive amount of data, the near-total lack of analysis makes it a sloppy report. Data does not equal understanding.
Mother Jones gets sloppy with jobs
From Mother Jones comes 12 Charts that Will Make Your Blood Boil, including the chart shown below. True to its promise, it made my blood boil - with sheer sloppiness. Take a look:
A cursory glance reveals the intended barb: greedy multinationals have caused us to create a ton of jobs overseas, and lose even more jobs here in the US!
But wait a second: is it possible that the US economy didn’t create a single domestic job in 10 years, between 2000 and 2009? All that red ink on the negative side seems to suggest so. But that’s obviously not the case.
- Are these figures only from the job gains and losses from multinationals, or are all companies included? What about jobs in the public sector? It’s never stated.
- What does it mean that “figures are cumulative since 2000” - every year’s number is the sum of all previous years? Huh?
- Are these raw numbers or net gains and losses? Again, never stated.
- Does it really help to begin a chart title with “Dude”?
If there’s actual data available about job losses due to outsourcing, Mother Jones readers deserve to see it. (Lots of other readers do, too.) Throwing up a sloppy chart like this discredits the entire conversation. Mother Jones should do much better.
Huffington Post gets sloppy with tax-cut data
A Huffington Post article describing “how Bush-era tax cuts changed America” features several charts from cbpp.org, including these two:
Do you notice anything missing? Other than a label for the x-axis, that is. (Probably a safe assumption that this is “annual household income.”) Anything else missing?
Well.. we see a distinct difference between those with $40,000 to $50,000 income, and those with over $1 million in income. And as for the other income levels… what happened to them?
Obviously: why doesn’t the chart mention the households earning between $51,000 and $1 million? And the households earning under $40,000? Given that the median US household income is around $50,000 (source) there are lots of households under $40k, and just over $50k.
Unless the chart shows tax-cut data for the full range of household incomes, it doesn’t impart much understanding. No matter where readers sit on the political spectrum, they should demand better charts than this. This is just… sloppy.
Silicon Valley hiring, charted sloppily
From this Top Prospect blog post, we learn that Silicon Valley companies are hiring each other’s employees. Take a look at the accompanying graphic, below:
The graph is visually impressive, with the colors and the cool-looking pattern created by the networked arrows, but as for imparting understanding to the reader, it’s pretty sloppy.
- important data variance is hidden: for example, the “2 to 1” arrow from LinkedIn to Facebook is the same size as the “11 to 1” arrow from Apple to Facebook.
- the most distinctive difference between arrows is the color, which is unnecessary due to the arrows already showing connectivity.
Now, a quick exercise for the reader. How do you explain the strange result that struggling Yahoo has a 10.5 multiple to Facebook, while Apple (rocking) has a nearly identical 11 multiple?
One immediate hypothesis is that the 10.5 multiple reflects how many people were hired but not how many applied in the first place. If a thousand Yahoo employees applied for 10 eventual hires, and a dozen Apple employees applied for 10 eventual hires, we’d see multiples about like the graph shows. Since the denominator is missing from the graph, we’re unable to get the full story.
NYT Magazine gets sloppy with life expectancy
I was amused to see this graph in the New York Times Magazine, showing the life expectancy for male and female 45-year-olds:
Instead of a bar graph with two lines showing 78 and 82, we get…
- two three-dimensional boxes growing obliquely out of the horizon line
- an ambiguous visual comparison: are we looking at the boxes’ height, or volume?
- two tones of shading on the boxes’ sides (darker for men, lighter for women)
- unnecessary precision in the numbers: is “78.11 years” really teaching the reader more than “78”?
- labels (“Men” and “Women”) unnecessarily separated from the data, and attached to the ambiguously sized boxes by dotted lines.
Why would the New York Times run such a sloppy graph? My only guess is that saying “45-year-old men live to 78, while women live to 82” was not visually exciting enough, so someone decided to juice it up.