Optimising the aspect ratio

The graph in the infographic referred to in this earlier post has an unusually large aspect ratio (i.e. the ratio of its width to its height).

The aspect ratio we choose for our graph has a significant effect on the perception of the data within it so optimising the aspect ratio is very important.

Consider the same data in that infographic plotted using two radically different aspect ratios:

Big aspect ratio

 

In this version, the rate of increase looks steady but not particularly impressive for an entirely new category of products.

Contrast this with a version of the same chart using an aspect ratio of less than 1:

Small aspect ratio

 

In this case, the growth rate looks exceedingly impressive.

In graphs such as this, which involve the reader making slope judgements, we need to take care to avoid introducing unforeseen bias.  WS Cleveland calls this “the angle contamination of slope judgements” in his book “The Elements of Graphing Data”.

This is illustrated in the pair of graphs below that plot exactly the same data points but have radically different aspect ratios.

Line judgements

The ratio of the slope of line segment BC to the slope of line segment AB is exactly the same in both cases.  However, the difference of the angles of these line segments with the horizontal is greater in the graph on the right than in the graph on the left.  This makes the line segments in the left graph appear to have slopes that are closer than those in the right panel.

The implication of this is that it is best to avoid extreme aspect ratios in graphs and to maintain consistency of aspect ratio between different charts in the same dashboard or report wherever possible.  An aspect ratio greater than 1 but less than 2 works well for trend charts.

Leave a Reply