Creating a grouped bar plot with Seaborn

Question

I am trying to create a grouped bar graph using Seaborn but I am getting a bit lost in the weeds. I actually have it working but it does not feel like an elegant solution. Seaborn only seems to support clustered bar graphs when there is a binary option such as Male/Female. (https://seaborn.pydata.org/examples/grouped_barplot.html)

It does not feel right having to fall back onto matplotlib so much - using the subplots feels a bit dirty :). Is there a way of handling this completely in Seaborn?

Thanks,

Andrew

import pandas as pd
import seaborn as sns
import matplotlib.pyplot as plt
from matplotlib import rcParams
sns.set_theme(style="whitegrid")
rcParams.update({'figure.autolayout': True})

dataframe = pd.read_csv("https://raw.githubusercontent.com/mooperd/uk-towns/master/uk-towns-sample.csv")

dataframe = dataframe.groupby(['nuts_region']).agg({'elevation': ['mean', 'max', 'min'],
                                                   'nuts_region': 'size'}).reset_index()
dataframe.columns = list(map('_'.join, dataframe.columns.values))

# We need to melt our dataframe down into a long format.
tidy = dataframe.melt(id_vars='nuts_region_').rename(columns=str.title)

# Create a subplot. A Subplot makes it convenient to create common layouts of subplots.
# https://matplotlib.org/3.3.3/api/_as_gen/matplotlib.pyplot.subplots.html
fig, ax1 = plt.subplots(figsize=(6, 6))

# https://stackoverflow.com/questions/40877135/plotting-two-columns-of-dataframe-in-seaborn
g = sns.barplot(x='Nuts_Region_', y='Value', hue='Variable', data=tidy, ax=ax1)

plt.tight_layout()
plt.xticks(rotation=45, ha="right")
plt.show()

Answer 1

I'm not sure why you need seaborn. Your data is wide format, so pandas does it pretty well without the need for melting:

from matplotlib import rcParams
sns.set(style="whitegrid")
rcParams.update({'figure.autolayout': True})
fig, ax1 = plt.subplots(figsize=(12,6))

dataframe.plot.bar(x='nuts_region_', ax=ax1)
plt.tight_layout()
plt.xticks(rotation=45, ha="right")
plt.show()

Output: