pandas boxplot by names of day of week: how to re-order the day-of-week names?

Using pandas boxplot(by="name_of_week_day"), I’d like to control the order of names of week days display on the plot. They are displayed in a wrong order, and I want them to be displayed in the order of the week, beginning by Monday and finishing by Sunday.

Here is a simple reproductible example:

import pandas as pd

#create DataFrame
df = (
    # we create a 2 columns df: date and sales
    pd.DataFrame({'date': pd.date_range(start='1/5/2022', freq='D', periods=15),
                   'sales': [6, 8, 9, 5, 4, 8, 8, 3, 5, 9, 8, 3, 4, 7, 7]})
    # we create a new column to get the name of the day of the week
    .assign(name_of_day = lambda df: df.date.dt.day_name())
    )

The above code delivers a df. Here are its first 3 rows:

date    sales   name_of_day
0   2022-01-05  6   Wednesday
1   2022-01-06  8   Thursday
2   2022-01-07  9   Friday

Now plot the boxplot with:

df.boxplot(by="name_of_day");

It returns the plot:

I would like the plot to deliver the names of the days of week in the right order.

How to do with pandas.boxplot() or with (pandas.plot.box() ?

Nota: yes, we could do it with seaborn, but my question is about pandas boxplot() or pandas plot.box().

>Solution :

Use a CategoricalDtype to reorder the days:

from calendar import day_name

days = pd.CategoricalDtype(list(day_name), ordered=True)

df.astype({'name_of_day': days}).boxplot(by="name_of_day")

Or in your original pipeline:

df = (
    # we create a 2 columns df: date and sales
    pd.DataFrame({'date': pd.date_range(start='1/5/2022', freq='D', periods=15),
                   'sales': [6, 8, 9, 5, 4, 8, 8, 3, 5, 9, 8, 3, 4, 7, 7]})
    # we create a new column to get the name of the day of the week
    .assign(name_of_day = lambda df: pd.Categorical(df.date.dt.day_name(),
                                                    categories=list(day_name),
                                                    ordered=True)
                                                   )
    )

Output: