Skip to main content

Critique: NYT unemployment graphics

To illustrate the horrible unemployment numbers announced today ("employers in the United States have shed about 3.6 million jobs since December 2007"), the New York Times presents a series of graphics:


We have a perfectly nice bar chart here: clear, legible, professionally done, all that. This chart tells us two (2) things: what month it is and how many jobs were lost. Surely there is additional info about this two-year stretch that would add depth and meaning (and potentially even yield an epiphany or two!).

One suggestion: The story says, "Manufacturers eliminated 207,000 jobs, more than in any year since 1982. The construction industry eliminated 111,000 jobs. And retailers, who were wrapping up their worst holiday shopping season in years, eliminated 45,000 jobs." Segmenting each bar by color to represent the various sectors would show at a glance where the biggest losses were coming from.


Again, a perfectly nice fever chart. (I don't mean to minimize the strong points of these two charts; they're clear, junk-free and, I'm sure, accurate. But that's why these folks got hired by the NYT and that's what they get paid for.) Thing is, there's much more that could be done here. What can we learn from this, aside from how high the unemployment rate has been over the last 20 years? Is there any other data set a designer could add to make this chart richer and more meaningful?

Another bit from the story: "'This is a horror show we’re watching,' said Lawrence Mishel, president of the Economic Policy Institute, an economic research organization in Washington. 'By every measure available — loss of employment and hours, rise of unemployment, shrinkage of the employment to population rate — this recession is steeper than any recession of the last 40 years, including the harsh recession of the early 1980s.'" [Note: the second sentence quoted here was removed from later versions of the story.]

Those all sound like interesting measures. Why not incorporate a series of sparklines to compare the trajectories of each of these trends with that of the unemployment numbers?


This timeline/bar chart, as you see, shows how unemployment rates have grown among various demographic groups. On the Times's site, the chart above is animated—we watch the arrows move from left to right. Fine, as far as it goes.

But it raises several questions for me: Why does this have to be animated (aside from making the project more multimedia-rific)? What purpose does the motion serve? And what is the point (no pun intended) of the arrowheads? Couldn't this graphic be static and simplified even further without losing meaning? In fact, if you simplify (not dumb down — there's a big difference) wherever you can, you then have more visual/conceptual scope for plotting additional data points... aka context.

The story itself lists dimensions that could be added to make these graphics truly masterful: "the country is trapped in a vortex of plunging consumer demand, rising joblessness and a deepening crisis in the banking system."

Stay tuned; in upcoming weeks I'll post my reworkings of these graphs to accord with my recommendations above.

Bureau of Labor Statistics report here.


  1. Chart 3 would be much more informative if it retained the date axis. We could then see how closely each of the four segments followed the trends of the overall economy in chart 2. There would be no need for mediatastic animation, nor even for arrowheads.

  2. That's so true, Jon. In fact, they could have obtained the figures in Chart 3 for each year back to 1990, and then plotted those as additional fever lines in various colors on Chart 2; that would show how each unemployment spike affects the demographic groups to different degrees.


Post a Comment

Popular posts from this blog

Cool Map: Emissions worldwide

From the Center for Public Integrity comes this interactive map showing greenhouse gas emissions from many of the world's largest economies.

Lots of data is packed into this simple interface, and the map itself is blessedly clear.

Note, though, that when it comes to infoviz issues, even these pros needed a do-over. Check out the message in the lower left corner. In an earlier version they made the common mistake of comparing circles based on radius, instead of by area. It's to their credit that not only did they fix the mistake, but they also owned up to it and made the change. The larger problem, though, is that distinguishing the relative size of circles is not easy for the average viewer; rectangles are clearer, and would probably have made this cool map even stronger.

Note also that stats are from 2005. Since then there's been substantial economic growth in China (for example), so the current numbers are likely to be even higher than what's shown here.

Differing Vie…

Quality dataviz about quality-of-life issues

To accompany its Better Life Initiative, OECD (the Organization for Economic Co-operation and Development) has put up a handsome, carefully constructed set of interactive data graphics called the Better Life Index:
There is more to life than the cold numbers of GDP and economic statistics – this Index allows you to compare well-being across countries, based on 11 topics the OECD has identified as essential, in the areas of material living conditions and quality of life. Each country is represented by a multicolored flower with 11 petals (OK, yes, potentially cheesy). The length of each petal represents the country's score in a given area; the width of the petal indicates the importance the user has assigned to that particular aspect. Drilling down into the details is easy to do; in fact, if you've a mind to do your own visualizations of this info, the underlying index data can be downloaded in spreadsheet format.

Kudos to Moritz Stefaner, Jonas Leist and Timm Kekeritz (for Raure…

Hans Rosling and Gapminder

As part of his larger mission of promoting "fact-based" public health policy, Swedish physician Hans Rosling founded, which aims to make world health data available and understandable to everyone. Back in 2006, Rosling gave a well-received TED presentation on the principles of Gapminder, showing, among other things, relative historical changes in life expectancy and GDP. (He spoke again in 2007.)

Rosling acknowledges that there are some small flaws and inconsistencies with data derived from all these different sources, but believes that the comparative results are far more significant. Check out the vast difference between Mauritius and Congo in income per person and life expectancy (shown on the Gapminder site and in the '06 presentation); consequently, says Rosling, using the term "sub-Saharan Africa" to describe both of these countries is vague to the point of uselessness.

His point: The more easily data and details can be visualized and compared…