Pandas: How to Create Bar Chart to Visualize Top 10 Values

You can use the following basic syntax to create a bar chart in pandas that includes only the top 10 most frequently occurring values in a specific column:

import pandas as pd
import matplotlib.pyplot as plt

#find values with top 10 occurrences in 'my_column'
top_10 = (df['my_column'].value_counts()).iloc[:10]

#create bar chart to visualize top 10 values
top_10.plot(kind='bar')

The following example shows how to use this syntax in practice.

Example: Create Bar Chart in Pandas to Visualize Top 10 Values

Suppose we have the following pandas DataFrame that contains information on the team name and points scored by 500 different basketball players:

import pandas as pd
import numpy as np
from string import ascii_uppercase
import random
from random import choice

#make this example reproducible
random.seed(1)
np.random.seed(1)

#create DataFrame
df = pd.DataFrame({'team': [choice(ascii_uppercase) for _ in range(500)],
                   'points': np.random.uniform(0, 20, 500)})

#view first five rows of DataFrame
print(df.head())

  team     points
0    E   8.340440
1    S  14.406490
2    Z   0.002287
3    Y   6.046651
4    C   2.935118

We can use the following syntax to create a bar chart that displays the top 10 most frequently occurring values in the team column:

import matplotlib.pyplot as plt

#find teams with top 10 occurrences
top_10_teams = (df['team'].value_counts()).[:10]

#create bar chart of top 10 teams
top_10_teams.plot(kind='bar')

The bar chart only contains the names of the top 10 most frequently occurring teams.

The x-axis displays the team name and the y-axis displays the frequency.

Note that we can also customize the plot to make it more aesthetically pleasing:

import matplotlib.pyplot as plt

#find teams with top 10 occurrences
top_10_teams = (df['team'].value_counts()).[:10]

#create bar chart of top 10 teams
top_10_teams.plot(kind='bar', edgecolor='black', rot=0)

#add axis labels
plt.xlabel('Team')
plt.ylabel('Frequency')

Note that the edgecolor argument added a black border around each bar and the rot argument rotated the x-axis labels 90 degrees to make them easier to read.