In this answer, we can see that pd.cut() results in a column of type category and that .apply() allows access to the properties of a given interval value.
For example, this works:
df['b'] = df['a'].apply(lambda x: x.left)
But this does not work:
df['b'] = df['a'].left
I don't understand why this is. Couldn't find a good explanation. If someone can explain, that would be great.
AmineBTG's comment on the question answers it.
In the following, .left applies to the series and there is no left method (or property) for pd.series
df['b'] = df['a'].left
In the following, .apply() works row-wise and acts on the series values one at a time. So .left is correctly interpreted as a property of each [interval] value.
df['b'] = df['a'].apply(lambda x: x.left)