My readers did it again! They submitted 28 entries of ways to visualize the table in Figure 1. This part of the Graph Makeover Contest was much more challenging than the redesign of the donut chart, but readers suggested a variety of ways to visualize the data in the table.
Figure 1 shows the table of data we want to visualize.
Of the 28 entries, twelve were variations of bar charts, seven were arrow charts, three were dot plots, three were slope charts and others were original figures without a standard name.
Before discussing the entries, it is worth noting that the table itself had a number of problems which made it difficult for readers to interpret the data. Some of the items are main categories (“Food”) and some are subcategories (“Food at home,” “Food away from home”). The total annual expenditures is equal to the sum of the main categories only (in this particular case subcategories cannot be used since they are not comprehensive). However, the difference between categories and subcategories was lost on some readers since every other line was highlighted, without regard to this important distinction (See Should I Shade Alternate Rows in My Tables? for a previous post on this problem.)
The problem was compounded by the fact that subcategories were only indented by one space. Creating accurate data visualizations is definitely more challenging when the format in which we receive the data is not clear. Good designers must clarify any ambiguities before attempting to visualize the data, as these readers did.
In addition, the table does not include a lot of interesting information. It would be useful to have data on other demographic variables to relate to expenditures. However, as you will see as you read this, the stories these data tell are much clearer from all of these entries than from the table itself.
As I noted in my last post, I invited both Alberto Cairo and Richard Heiberger to comment on some of the figures to give those who submitted entries more than one opinion. Alberto Cairo teaches at the University of Miami and Richard Heiberger teaches at Temple University. Once again, as Alberto pointed out, many of the comments that follow are opinions since there have not been rigorous studies of many aspects of graphing. Many thanks to both Alberto and Rich for their very useful comments.
Please comment on the article here: Effective Graphs