Bar graph/chart is a common type of visualization used to present categorical data using rectangular bars. Bar charts come in different types such as vertical, horizontal, stacked (either vertical or horizontal), grouped and 100% stacked bar charts. It is used to compare the relative sizes between two or more categories of data. In this post we will look at Bar Graphs and how to create Bar Graph With Matplotlib.
When to Use Bar Graph
- Compare between two or more different groups in data.
- Track change over time. Note: Suitable when the changes are too large otherwise use line graph.
How to Use Bar Graph
- Bars should be of same type. Avoid mixing 3D and 2-D bars.
- The scale of values should always start from zero unless otherwise.
- Select the colours to use carefully.
- Order the bars with some criteria that best presents the insights.
- To maximise the visibility and appearance of all components use different variations of bar charts (vertical, horizontal, grouped etc.).
- Include values in the bars for users to make actual comparison between groups.
Bar Graph in Matplotlib
Create DataFrame
# Load required libraries
import matplotlib.pyplot as plt
import pandas as pd
import numpy as np
plt.style.use(‘ggplot’) # Other sytles to use; fivethirtyeight
score_df = pd.DataFrame(
{
"Students": ["Tom", "Peter","Simon", "Mary", "Jane","King","Hillary","Ethan","Page"],
"Math": [79.00, 67.00,80.00, 84.00, 70.00,60.00,90.00,76.00,75],
"Physics":[63.00, 98, 60.00, 90,84.00, 77.00,55.00,70,66.00],
"Computer":[84.00,78.00, 57.00, 88.00, 75.00,93.00,92.00,98.00,90.00],
},
index=["Tom", "Peter","Simon", "Mary", "Jane","King","Hillary","Ethan","Page"]
)
score_df['Total']=score_df[['Math','Physics','Computer']].apply(np.sum,axis=1)
score_df
Simple Bar Graph
plt.figure(figsize=(20,10)) # Set figure size
plt.bar(score_df['Students'],score_df['Total'],color='skyblue')
plt.title('Students Score',fontsize=24)
plt.xticks(rotation=45)
plt.xlabel('Students',fontsize=24)
plt.ylabel('Score',fontsize=24)
plt.show()
Add Labels
def addlabels(x,y):
for i in range(len(x)):
plt.text(i, y[i], y[i], ha = 'center')
plt.figure(figsize=(20,10)) # Set figure size
plt.bar(score_df['Students'],score_df['Total'],color='skyblue')
addlabels(score_df['Students'],score_df['Total'])
plt.title('Students Score',fontsize=24)
plt.xticks(rotation=45)
plt.xlabel('Students',fontsize=24)
plt.ylabel('Score',fontsize=24)
plt.show()
Add Labels at the middle
def addlabels(x,y):
for i in range(len(x)):
plt.text(i, y[i]//2, y[i], ha = 'center')
plt.figure(figsize=(20,10)) # Set figure size
plt.bar(score_df['Students'],score_df['Total'],color='skyblue')
addlabels(score_df['Students'],score_df['Total'])
plt.title('Students Score',fontsize=24)
plt.xticks(rotation=45)
plt.xlabel('Students',fontsize=24)
plt.ylabel('Score',fontsize=24)
plt.show()
Horizontal Simple Bar Graph
plt.figure(figsize=(20,15)) # Set figure size
plt.barh(score_df['Students'],score_df['Total'],color='skyblue')
plt.title('Students Score',fontsize=24)
plt.xticks(rotation=45)
plt.xlabel('Students',fontsize=24)
plt.ylabel('Score',fontsize=24)
plt.show()
Grouped Bar Chart
score_df[['Math','Physics','Computer']].plot(kind="bar",figsize=(20, 10))
plt.title('Students Score',fontsize=24)
plt.xticks(rotation=45)
plt.xlabel('Students',fontsize=24)
plt.ylabel('Score',fontsize=24)
plt.show()
Stacked Bar Chart
score_df[['Math','Physics','Computer']].plot(kind="bar",stacked=True,figsize=(20, 10))
plt.title('Students Score',fontsize=24)
plt.xticks(rotation=45)
plt.xlabel('Students',fontsize=24)
plt.ylabel('Score',fontsize=24)
plt.show()
For complete code check the jupyter notebook here.
Conclusion
Bar charts are important type of visualizations for presenting data with discrete groups. They are used to compare relative relationship between categories with bars. They are also useful in showing trend between large time intervals. In this post we have looked at what’s bar chart, when to use them and how to use them with Matplotlib. In the next post we will look at Pie Charts and how to use them in Matplotlib. To learn about line graph check our previous post here.