Last week, I briefly mentioned my concern that New York Times’ new VisLab would discourage the Times from creating their lovely and frequently enlightening visualizations by providing an easy way to crowd-source visualizations. Since the Times’s visualizations are often one-offs and customized for the data at hand, I referred to them as “couture” visualizations, as opposed to the “ready-to-wear” options of Excel, Many Eyes, Swivel, and other similar services.

Over at Uncertain Principles, Chad Orzel provides a great example of how a custom visualization can be far superior to a “ready made” one. Orzel blogs about the problems he sees with the default Excel graph comparing Obama’s and McCain’s tax proposals:

His analysis of the problems with the graph is dead on. The graph displays percent change in tax by income quintile, or as Orzel puts it: “the graph-maker has plotted confusion as a function of obfuscation”.

The Washington Post provides a better visualization:

As Orzel notes, this graph plots the change in taxes as a dollar amount, against incomes, again, shown in a dollar amount. All of a sudden, it becomes obvious that the big difference is for those families making over $603,403 per year. However, while dollar amounts are shown, the y-axis is fairly meaningless and information about the number of people in each tax bracket is missing, perhaps leading readers to conclude that there is an equal number of individuals in the “above $2.87 million” bracket as in the “below $18,981” bracket.

A commenter on Orzel’s blog suggested yet another version:

Now, both the x- and y- axes are meaningful, where the y-axis now represents population, and additional information regarding median household income and median individual income has been added.

While I appreciate the quick prototyping and analysis that Excel’s charting functions allow one to do, the custom visualization contains important details that significantly change one’s understanding of the situation. Developing a “couture” visualization takes time, skill, and attention to detail, all of which can significantly increase the value and the impact of the visualization. Sometimes it’s worth the extra investment.