Splits data into groups by a categorical dimension.
Each subgroup of the data is contained in a
<group>, which has scope properties
data (corresponding to the data subset) and
group (the value of the grouping key for that subgroup).
The dimenson of the data over which to split the data into subgroups.