[Python] Order of columns in a plotnine bar plot using a polars dataframe

I'm quite new to the packages polars and plotnine and have the following code:

import polars as pl
import polars.selectors as cs
from plotnine import *

df = pl.read_csv('https://raw.githubusercontent.com/Brinkhuis/Zorguitgaven/refs/heads/master/zorguitgaven.csv')
df = df.unpivot(cs.numeric(), index='Category')

(
ggplot() +
geom_bar(
data=df.filter(pl.col('variable').is_in(['Mannen 2040', 'Vrouwen 2040'])),
mapping=aes(x='Category', y='value', fill='variable'),
stat='identity'
) +
scale_fill_manual(values=['#007BC7', '#CA005D']) +
coord_flip()
)


It runs without errors. However, the Category order is not right.



The "5-10" value show be one "row 3".

My questions:

1. Is there a way to set the categorical order in a polars dataframe to fix this?
2. Is there a way to set the categorical order in plotnine?

Continue reading...