How to annotate countplot with percentages by category

Question

Hi I'm trying to add percentages to my countplot with 5 categories and 2 values (old and younger). I've tried adding the def and loop from How to add percentages on top of bars in seaborn?

My code:

plt.figure(figsize =(7,5))
ax = sb.countplot(data = df_x_1, x = 'concern_virus', hue = 'age')
plt.xticks(size =12)
plt.xlabel('Level of Concern', size = 14)
plt.yticks(size = 12)
plt.ylabel('Number of People', size = 12)
plt.title("Older and Younger People's Concern over the Virus", size = 16)
ax.set_xticklabels(ax.get_xticklabels(), rotation=40, ha="right");

for p in ax.patches:
    percentage = '{:.1f}%'.format(100 * p.get_height()/total)
    x = p.get_x() + p.get_width()
    y = p.get_height()
    ax.annotate(percentage, (x, y),ha='center')
plt.show()

As you can see, the percentages don't make sense.

Answer 1

The problem seems to be with the variable that is undefined in the above code: total. total should be the number you want to call 100%, for example the total number of rows in the dataframe. That way all the displayed percentages sum up to 100.

Here is some sample code:

import matplotlib.pyplot as plt
import pandas as pd
import numpy as np
import seaborn as sns

N = 250
df_x_1 = pd.DataFrame({'concern_virus': np.random.choice(['a', 'b', 'c', 'd', 'e'], N),
                       'age': np.random.choice(['younger', 'older'], N)})
plt.figure(figsize=(7, 5))
ax = sns.countplot(data=df_x_1, x='concern_virus', order=['a', 'b', 'c', 'd', 'e'],
                   hue='age', hue_order=['younger', 'older'],
                   palette=['chartreuse', 'darkviolet'])
plt.xticks(size=12)
plt.xlabel('Level of Concern', size=14)
plt.yticks(size=12)
plt.ylabel('Number of People', size=12)
plt.title("Older and Younger People's Concern over the Virus", size=16)
ax.set_xticklabels(ax.get_xticklabels(), rotation=40, ha="right")

total = len(df_x_1)
for p in ax.patches:
    percentage = f'{100 * p.get_height() / total:.1f}%\n'
    x = p.get_x() + p.get_width() / 2
    y = p.get_height()
    ax.annotate(percentage, (x, y), ha='center', va='center')
plt.tight_layout()
plt.show()

To have the text in the center of the bar, it helps to choose ha='center' and add half the width to the x-position. Appending a newline to the text can help to position the text nicely on top of the bar. plt.tight_layout() can help to fit all the labels into the plot.

Seaborn lets you fix the order of the x-axis via order=.... The order of the legend elements and the corresponding colors can be set via hue_order=... and palette=....

PS: For the new question, with totals per age group, instead of directly looping through all the bars, a first loop can visit the groups:

import matplotlib.pyplot as plt
import pandas as pd
import numpy as np
import seaborn as sns

label_younger = 'younger'
label_older = 'older'
df_younger = pd.DataFrame({'concern_virus': np.random.choice(['a', 'b', 'c', 'd', 'e'], 230)})
df_older = pd.DataFrame({'concern_virus': np.random.choice(['a', 'b', 'c', 'd', 'e'], 120)})
df_younger['age'] = label_younger
df_older['age'] = label_older
df_x_1 = pd.concat([df_younger, df_older], ignore_index=True)

plt.figure(figsize=(7, 5))
ax = sns.countplot(data=df_x_1, x='concern_virus', order=['a', 'b', 'c', 'd', 'e'],
                   hue='age', hue_order=[label_younger, label_older],
                   palette=['orangered', 'skyblue'])
plt.xticks(size=12)
plt.xlabel('Level of Concern', size=14)
plt.yticks(size=12)
plt.ylabel('Number of People', size=12)
plt.title("Older and Younger People's Concern over the Virus", size=16)
ax.set_xticklabels(ax.get_xticklabels(), rotation=40, ha="right")

for bars in ax.containers:
    if bars.get_label() == label_younger:
        group_total = len(df_younger)
    else:
        group_total = len(df_older)
    for p in bars.patches:
        # print(p.get_facecolor(), p.get_label())
        percentage = f'{100 * p.get_height() / group_total:.1f}%\n'
        x = p.get_x() + p.get_width() / 2
        y = p.get_height()
        ax.annotate(percentage, (x, y), ha='center', va='center')
plt.tight_layout()
plt.show()

Answer 2

• As per How to add percentages on top of grouped bars and How to add value labels on a bar chart, this can be more easily accomplished with matplotlib.pyplot.bar_label
• Tested in python 3.11.2, pandas 2.0.0, matplotlib 3.7.1, seaborn 0.12.2

Imports and DataFrame

import seaborn as sns
import pandas as pd
import numpy as np

np.random.seed(2023)
N = 250
df = pd.DataFrame({'concern_virus': np.random.choice(['a', 'b', 'c', 'd', 'e'], N),
                   'age': np.random.choice(['younger', 'older'], N)})

# get the total count of the entire age column
total = df.age.count()

# get the count of each group in the age column
vc = df.age.value_counts()

Plot as percent of age count total

ax = sns.countplot(data=df, x='concern_virus', hue='age', order=sorted(df.concern_virus.unique()))
sns.move_legend(ax, bbox_to_anchor=(1, 0.5), loc='upper left', frameon=False)

for c in ax.containers:
    ax.bar_label(c, fmt=lambda v: f'{(v/total)*100:0.1f}%')

Plot as percent of age group count total

ax = sns.countplot(data=df, x='concern_virus', hue='age', order=sorted(df.concern_virus.unique()))
sns.move_legend(ax, bbox_to_anchor=(1, 0.5), loc='upper left', frameon=False)

for c in ax.containers:
    label = c.get_label()
    total = vc[label]
    ax.bar_label(c, fmt=lambda v: f'{(v/total)*100:0.1f}%')