26 Chart scales
By now, I hope it is clear that when making a visualisation in ggplot, we do so by a process of mapping data to aesthetics. To give a concrete example, when we write x = GDPperCAP, y = lifeExp, color = continent, behind the scences ggplot is taking the values in the columns we give it, and giving them some kind of visual form. In the case of the position (x and y), by deftault this means that distance on the plot corresponds to equal amounts numerical values. In the case of colour, ggplot2 knows there are five distinct values, and that it should give each a distinct colour.
Scales
The process by which this happens involve scales. Scales are the next ‘layer’ in our ggplot object. Scales control the mapping from data to aesthetics. Data visualisations cannot work without scales, so ggplot2 will automatically add them to your plots, but knowing how to control these and the various options they have available will allow you to have much more control over the look of your visualisations.
These scales are communicated to the viewer by means of an axis or a legend, which we’ll cover afterwards.
We won’t go into the theory of how g gplot2 works with scales in too much detail, but concentrate on some practical code, examples and tips which I think have the most impact on controlling the aspects of your plots you would expect to be able to do.
Aesthetics
Recall from last week that we learned about different aesthetics: position, colour, size, shape, linetype, linewidth, and transparency. This chapter on scales will mostly cover position and colour, and to a lesser extent, size. The scales for the other aesthetics are less interesting (you’re less likely to want to change the defaults), so we won’t look at these. But be aware they also have scales which can be overwritten or adjusted.