How to plot multiple pie chart or bar chart from multi-index dataframe?

Question

Dataframe is as below.

There are multiple index as category_1 & category_2 and two column as count & %

I need to draw multiple (for each category_1) pie chart for category_2 & % and bar chart for category_2 & count.

		count	%
category_1	category_2
Bady and Kids	Baby Care	220	27.5
	Toys	176	22
	Boys Clothing	156	19.5
	Girls Clothing	144	18
	Baby Boy Clothing	104	13
Women’s wear	Party Dresses	224	28
	Sports Wear	188	23.5
	Swim wear	140	17.5
	Winter Wear	128	16
	Watches	120	15

Asked By: Priyank Patel

||

Source

Answer 1

You can use:

#gt number of groups first for subplots
no = len(df.index.levels[0])
fig, axes = plt.subplots(no, 2, figsize=(10, 6))

#for each group by first level of Multiindex
for i, (name, g) in enumerate(df.groupby(level=0)):
    #remove first level and index name
    g = g.droplevel(0).rename_axis(None)

    #get positions dynamic
    ax =  axes[i, 0]
    ax1 = axes[i, 1]

    #plot both graphs together
    g['count'].plot.pie(ax=ax, legend=False)
    ax.set_title(f'{name} %')
    g['%'].plot.bar(ax=ax1, legend=False, rot=45)
    ax1.set_title(f'{name} count')

fig.subplots_adjust(wspace=.2)

Answered By: jezrael

Answer 2

You can use the following code to plot pie charts and bar charts for each category_1 in your multi-index dataframe:

import pandas as pd
import matplotlib.pyplot as plt

# create a sample dataframe
data = {
    ('Bady and Kids', 'Baby Care'): (220, 27.5),
    ('Bady and Kids', 'Toys'): (176, 22),
    ('Bady and Kids', 'Boys Clothing'): (156, 19.5),
    ('Bady and Kids', 'Girls Clothing'): (144, 18),
    ('Bady and Kids', 'Baby Boy Clothing'): (104, 13),
    ("Women's wear", 'Party Dresses'): (224, 28),
    ("Women's wear", 'Sports Wear'): (188, 23.5),
    ("Women's wear", 'Swim wear'): (140, 17.5),
    ("Women's wear", 'Winter Wear'): (128, 16),
    ("Women's wear", 'Watches'): (120, 15)
}
df = pd.DataFrame(data.values(), index=pd.MultiIndex.from_tuples(data.keys(), names=['category_1', 'category_2']), columns=['count', '%'])

# plot pie charts and bar charts for each category_1
for category_1 in df.index.get_level_values(0).unique():
    sub_df = df.loc[category_1]
    fig, (ax1, ax2) = plt.subplots(1, 2, figsize=(10, 5))
    sub_df.plot(kind='pie', y='%', legend=None, ax=ax1)
    ax1.set_title(f'{category_1} - %')
    sub_df.plot(kind='bar', y='count', legend=None, ax=ax2)
    ax2.set_title(f'{category_1} - count')
    plt.show()

This code creates a loop to iterate through each unique value of category_1. Inside the loop, it extracts a sub-dataframe for the current category_1 and then plots a pie chart for the % column and a bar chart for the count column using sub_df.plot(). Finally, it sets the title of each chart and shows them using plt.show().

This will output a set of pie charts and bar charts, one for each unique category_1.

Here is the sample charts:

Answered By: tarle

Answer 3

You can use:

for cat1, subdf in df.groupby(level='category_1'):
    fig, (ax1, ax2) = plt.subplots(1, 2, figsize=(12, 6))
    subdf.loc[cat1, :].plot.pie(y='count', ax=ax1, ylabel='', legend=False)
    subdf.loc[cat1, :].plot.bar(y='%', ax=ax2, rot=45, xlabel='', legend=False)
    fig.suptitle(cat1)
    fig.tight_layout()
    plt.show()

Output:

Answered By: Corralien

How to plot multiple pie chart or bar chart from multi-index dataframe?

Question:

Answers: