Re-doing the Average Annual Expenditures
The post "Visualizing average annual expenditures" was devoted to visualize the second example of the Forbes Graph Makeover Contest, a table with data about average annual expenditures by categories of consuming expending.
See bellow the first version of my entry to the contest:
In the article "Visualizing the Table of the Graph makeover Contest", Naomi Robbins presented the results of the contest with valuable comments, critiques and recommendations to seven entries, which showed the variety of graph types submitted and posts that showed solutions to others that had a problem. She also higlighted that there is more than one way to plot a data set.
Regarding the table itself she mentioned:
"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 it is received is not clear. Good designers must clarify any ambiguities before attempting to visualize the data, as these readers did."
Regarding my entry, Naomi Robbins gave the following recommendations:
"Shading every other row is problematic for the reasons discussed above. We would rather see color used to distinguish homeowners from renters than duplicating the information in the length and direction of the bars to show the percent change as shown in the color legend. That would free up the housing tenure column and allow for longer bars in the expenditures column. As it is, the lines for small changes in percent change are longer than those for big differences in expenditures. Also, including the total expenditures in the chart hurts the resolution of the other categories. The total annual expenditures could be mentioned in the title or a subtitle. I find that repeating K and % at every tick mark label clutters the labels and makes the numbers themselves more difficult to read. I would prefer the K and % signs to be in the axis labels."
Naomi also encouraged all authors to re-do their entries, by saying "I would be happy to posting before and after figures", so I take the challenge trying to do my best re-doing the dataviz following her recommendations. As a result I got the second version of my entry. See it below.
I think it is important to show the "Annual Expenditure" because it informs that spending in 1986 is pretty similar to 2010, however there are some categories and subcategories with significant decreasing and others with increasing from from 1986 to 2010.
I decided to add a new subcategory to the main categories Transportation and Healthcare to count for the rest in them. This allows to confirm the relevance of the subcategory "Gasoline and motor oil" in the category Transportation, as it is the case of "Health insurance" in Healthcare, both with a significant increase of expending in 2010 compared to 1986.
Key findings from this data visualization were already higlighted in the previous arcticle "Visualizing average annual expenditures"
As always, I appreciate very much your comments and recommendations regarding this new data visualization design.