Skip to content
Advertisement

python matplotlib “merge” x axis labels for multi index

I would like to “merge” the first level of multi index, i.e. have a centered “USA” and “EU” instead of the tupples.

Minimal Exmaple

df = pd.DataFrame(data={"region":["USA","USA","EU","EU"], "sctr":["HH","NFC","HH","NFC"], "values":[1,2,3,4]})
df = df.set_index(["region","sctr"])

fig, ax = plt.subplots(1,figsize=(8,6))
df.plot(kind="bar")
plt.xticks(rotation = 0)
plt.show()

Output enter image description here

Desired output Using excel, one gets the desired output (almost) by default:

enter image description here

The ideal solution is simple, short and does not require any manual adjustments if more regions / sectrs are added.

Advertisement

Answer

There are other ways to make the x-axis ticks into two lines: the first line is taken from the multi-index and made into a list; the second line is placed from the multi-index list by the text function. Hiding the existing labels is handled by the plot parameter.

import matplotlib.pyplot as plt
import pandas as pd

df = pd.DataFrame(data={"region":["USA","USA","EU","EU"], "sctr":["HH","NFC","HH","NFC"], "values":[1,2,3,4]})
df = df.set_index(["region","sctr"])

fig, ax = plt.subplots(1,figsize=(8,6))
df.plot(kind='bar', xlabel='', ax=ax)
ax.set_xticklabels(df.index.get_level_values(level=1).tolist(), rotation=0)

 for container in ax.containers:
    for i,child in enumerate(container.get_children()):
        if i == 0:
            ax.text(child.xy[0]+child.get_width(), -0.08, df.index.get_level_values(level=0)[0], ha='center', transform=ax.transAxes)
        elif i == 2:
            ax.text(child.xy[0]-(child.get_width()*2), -0.08, df.index.get_level_values(level=0)[2], ha='center', transform=ax.transAxes)

plt.show()

enter image description here

User contributions licensed under: CC BY-SA
3 People found this is helpful
Advertisement