Simulating individual order of columns within categories

This demo will show how you can simulate an order of columns from different ColumnSeries within each category.

It uses the following things to achieve it:

Each ColumnSeries has clustered = false set so that it always takes up the whole width of the “hidden” categories.
Data is manipulated so that each category contains only one value for just a single series.
Regular axis labels are disabled.
Axis ranges are used to place a label/grid to “group” several “hidden” categories.