You may have met five principles for creating good graphs in a course or seminar given by the Statistical Consulting Centre. Show You can also read about them in this article: Gordon I & Finch S. (2015). Statistician heal thyself: have we lost the plot? Journal of Computational and Graphical Statistics, 24(4), 1210-1229. Here are the principles and some tips for creating graphs that are consistent with them. Show the data clearlyShowing the data clearly includes ensuring the data points can be seen but also providing meaningful text on the graph itself. Tips include:
Use simplicity in design of the graphA graph with a simple design strives for a clean, uncluttered look. Simplicity in design does not mean simplicity in data however; well-designed graphs can represent rich data.
Use alignment on a common scaleGood graphs support accurate estimation of the quantities represented. To estimate quantities, the reader needs to understand the scale used to represent quantity on the graph.
The 2016 popular vote in the US Presidential election provides a simple example of a transposition: Transposition makes the labels easier to read here, and provides more landscape for the quantitative axis, making the quantities easier to estimate. Keep the visual encoding transparentGraphs use visual encoding to represent data. Readers need to decode the graph. This works best when the decoding task is made easy by astute choices in the design of the graph, making the visual encoding transparent and the decoding effortless. Here are some tips:
Use standard forms that workAccurate interpretation of graphs relies on a shared understanding of the definition of the graphical objects used. Using standard forms for standard needs supports more reliable interpretation of data visualisations.
|