Assignment 2: Chartjunk
What is Chartjunk?
According to Edward R. Tufte, the three main types of chartjunk are as follows:
- Moiré vibrations — unintentional optical art
- Grids — for initial plotting of data, but not for printing
- Ducks — focus on graphical style rather than data
In the data graphic above, we see four related sets of data concerning the consumption of metals in the United States in 2006. The bar chart (1) to the left shows the 2006 US consumption per capita in grams (although some quantities are written in kilograms) of 19 different metals. The measurements in grams are noted on the Y axis, while the different metals are labelled along the the X axis. Above each bar, the exact measurement in grams (or kilograms) is specified. In the middle of the data graphic, a stacked bar chart (2) is arranged in a circular fashion (resembling a polar area chart) with added icons symbolizing the use of each type of metal, in addition to a textual explanation. This chart shows the number of years left of metal consumption at today’s rate (2006) compared to the number of years left if the world were to consume at half the rate of the United States. Text in the center of the chart explains the arrangement of the data, as well as how this data was calculated and its limitations. To the right of this chart, 19 donut charts (3) are displayed at random, illustrating the proportion of consumption met by recycled materials in percentages. Both the sizes and the shading of the donut charts communicate the varying percentages of metal recycling. In the middle of each donut chart, the universal symbol for recycling is displayed, further illustrating the purpose of the data. Directly below the donut charts is a pictogram (4) depicting the world population in April 2007 versus the US population at the same point in time. The exact quantities are also specified on either side of the pictogram.
Based on the above principles outlined by Tufte, we can identify some elements of chartjunk within this data graphic:
Moiré vibrations
- The background of this data graphic is very busy. There are concentric circles in the middle (following chart 2), combined with horizontal lines on either side in varying colors.
- The large range of colors confuses the eye and complicates the data. 19 different colors are consistently used to identify the different types of metal, however this is unnecessary as the metals are labelled in all charts.
- Chart 4 in the bottom right corner has an optical art effect with the rows of white dots symbolizing heads of people. In addition to being a distraction to the eye, this chart is rather irrelevant as it merely supports the two population numbers, which are already clearly labelled.
Grids
- A type of grid is present behind the chart 2, in the form of concentric circles. Furthermore, the parallel lines on either side, forming the backgrounds of charts 1 and 3, resemble a type of grid as well, yet they do not help the viewer to read the data in the slightest.
Ducks
- Overall, this graphic elevates design over data, as it focuses on promoting stylistic elements (colors, shapes, backgrounds, icons) rather than communicating the quantitative information in an honest and direct manner.
In his book, The Visual Display of Quantitative Information, Tufte also describes a principles which he calls “data-ink”. He developed a formula by which the ratio of data-ink can be measured, i.e. the proportion of a graphic’s ink devoted to the display of quantitative information. In the data graphic above, there is much data-ink which can be erased, in the forms of non-data-ink and redundant data-ink.
Examples of non-data-ink:
- The horizontal lines and concentric circles forming the background for the entire graphic.
- The various little icons describing common uses of each type of metal in chart 2.
- The recycling symbols inside each of the 19 donut charts.
Examples of redundant data-ink:
- The 3D effect of the bars in chart 1.
- The 3D pictogram supporting the data in chart 4.
- The drop shadows in chart 3.
In addition to this, there is also a large amount of text in this graphic which could easily be reduced. Overall, according to Tufte’s principles, this data graphic has a low data-ink ratio, indicating that much of the ink used in this graphic is not directly devoted to displaying the quantitative information.
Below are improved versions of charts 1–3, according to Tufte’s principles. Chart 4 has not been redesigned as it does not really serve a purpose other than to support the two population quantities, which, in my opinion, speak for themselves.
