This section regroups several chart types used to study the joint distribution of 2 numerical variables. All these methods are really useful to avoid over plotting in a scatterplot. If you have too many dots, they basically counts the number of observations

within a particular area of the 2D space and represent this count by a color. If you divide the space by several squares you get a 2D histogram. If you use hexagons you get a hexbin plot. You can also calculate Density estimate and represent 2D density plots

or Contour plots. Fortunately, the ggplot2 library has awesome geoms to easily produce this kind of charts. Note that it is a good practice to show the marginal distributions of both variables.




2d histogram

This is the 2d version of the classic histogram. The plot area is split in a multitude of small squares, the number of points in each square is represented by its color.


This is the equivalent of the 2d histogram, but the plot area is split in a multitude of hexagones this time. This technic is also appreciated to create hexbin maps.


2d distribution

Like it is possible to plot a density instead of an histogram to represent a distribution, it is possible to make a 2d density plot. Several variations are available using ggplot2:


Without ggplot2

It is also possible to realise this kind of plot without ggplot2, even if it is often way more complicated..



Search the gallery