Points — Strip Plots

We’ll begin with the following data set:

\[x = (2.7, 3.5, 5.2, 5.6, 6.2, 6.3, 8.7, 9.8, 10.1, 10.4, 11.0, 12.1, 12.5, 12.5, 13.5, 14.5, 15.8, 17.4, 19.4)\]

Over-plotting is always a concern when using point and can be overcome by changing shape and transparency.

Strip plots are one of the easiest and simplest ways to visualise a distribution. Here, we encounter the frequently observed problem of over-plotting - multiple data points obscure each other and the distribution is falsely represented. To overcome this, we can use four methods:

Symbols
Such as the hollow circle in the second plot are an improvement, but do not eliminate over-plotting.
Jittering
Shown in the third plot, scatters the points so that they are not overlapping. It is only applied to the y-axis here. Alpha-blending {#Alpha-blending} or transparency, shown in the fourth plot, permits colour to reveal over-plotting. This is particularly useful as an alternative to jittering.
Size
In combination with alpha-blending in the fifth plot emphasises high-density areas of the data set.
Size
in combination with alpha-blending and jittering in the sixth plot goes further and allows you to see each point in a data set clearly.

Avoid over-plotting by using jittering and open circles for small data sets and alpha-blending for large data sets