Summary

This web content provides a comprehensive tutorial on creating a customized bar plot using Matplotlib and Seaborn in Python, aimed at enhancing the visual appeal and readability of data visualizations.

Abstract

The article titled "Draw a unique barplot using Matplotlib in Python" introduces Matplotlib as a fundamental and versatile library for data visualization in Python, complemented by Seaborn for statistical graphics. It guides readers through improving the aesthetics of a default bar plot by demonstrating steps to increase label font sizes, remove tick lines and chart borders, reorder bars for better comparison, change bar colors, add grid lines for precise data reading, and beautify the chart title and labels. The tutorial emphasizes the importance of data visualization quality and provides a step-by-step transformation of a basic bar chart into a visually appealing and informative graph, using minimum wage data across Canadian provinces as an example.

Opinions

The author suggests that the default bar plot generated by Seaborn may not be visually optimal and encourages customization for improved data presentation.
Enlarging font sizes and removing unnecessary elements like tick lines and borders are recommended for clearer communication of information.
The tutorial advocates for the use of color palettes to enhance the visual appeal of bar charts and suggests specific palettes like 'plasma' or shades of a single color.
Adding grid lines is presented as a method to assist the audience in accurately interpreting the data represented by each bar.
The author emphasizes the significance of a well-designed title and axis labels, proposing the use of background colors, larger fonts, and italicized text for emphasis.
The article concludes by showcasing the before and after comparison of the bar chart, highlighting the effectiveness of the applied enhancements in creating a more engaging and reader-friendly visualization.
The author offers additional resources for creating various types of charts, indicating a commitment to helping readers expand their data visualization skills beyond bar plots.

Draw a unique barplot using Matplotlib in Python

Matplotlib is the most basic and powerful library for data visualization in Python. It is a 2D plotting library that allows you to create publication-quality figures easily. Another complimentary package that is based on this data visualization library is Seaborn. It can provide a high-level interface to draw statistical graphics.

If you are tired of the bar plot generated by seaborn by default and interested in improving the quality and beauty of your bar chart, this tutorial is made for you.

In this tutorial, we are going to build a customized bar plot using Matplotlib and seaborn. The finished bar chart will look like this.

1. Import libraries

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

2. Create a dataframe for data visualization

Mydict={
 ‘Province’:[‘Alberta’,’British Columbia’,’Manitoba’,’New Brunswick’,’Newfoundland & Labrador’,
 ‘Northwest Territories’,’Nova Scotia’,’Nunavut’,’Ontario’,’Prince Edward Island’,
 ‘Quebec’,’Saskatchewan’,’Yukon’],
 ‘Province_Code’:[‘AB’,’BC’,’MB’,’NB’,’NL’,’NT’,’NS’,’NU’,’ON’,’PE’,’QC’,’SK’,’YT’],
 ‘Minimum_Wage’:[‘15.00’,’14.60',’11.65',’11.70',’11.65',’13.46',’12.55',’13.00',’14.00',’12.85',’13.10',’11.32',’13.71']
}

df=pd.DataFrame(Mydict)

#change the datatype of minimum wage from object to float
df['Minimum_Wage'] = df['Minimum_Wage'].astype(str).astype(float)
print(df.dtypes)

The dataset is very simple. It contains the minimum salaries across all provinces in Canada.

3. Draw a basic bar plot

fig,ax = plt.subplots(figsize=(9,6))
sns.barplot(x=’Province_Code’,y=’Minimum_Wage’,data=df,ci=95,ax=ax)
ax.set_title(‘Minimum Wage Comparison across Canada’)

It can be clearly seen that the font of the axis labels on the horizontal and vertical coordinates is a little bit small, It is hard to read. And the scale line is not very helpful in pinpointing the real number for each bar. The first step is to enlarge the font of the axis labels and remove the scale line.

3.1 Remove the tick line

fig,ax = plt.subplots(figsize=(9,6))
sns.barplot(x=’Province_Code’,y=’Minimum_Wage’,data=df,ci=95,ax=ax)
ax.set_title(‘Minimum Wage Comparison across Canada’)
 
# set the font size of labels into 16 and remove the tick line
ax.tick_params(labelsize=16,length=0)

3.2 Remove the borders of the bar chart

# method 1
ax.spines[‘left’].set_visible(False)
ax.spines[‘top’].set_visible(False)
ax.spines[‘right’].set_visible(False)
ax.spines[‘bottom’].set_visible(False)

#method 2
plt.box(False)

The chart is getting better already by increase fonts and removing the border lines.

3.3 Reorder the sequence of bars in the chart

fig,ax = plt.subplots(figsize=(9,6))
sns.barplot(x=’Province_Code’,y=’Minimum_Wage’,data=df,ci=95,ax=ax,
 order = df.sort_values(‘Minimum_Wage’).Province_Code, #reorder bars   ) #change the color of bars

ax.set_title(‘Minimum Wage Comparison across Canada’)
 
 
# set the font size of labels into 16 and remove the tick line
ax.tick_params(labelsize=16,length=0)
plt.box(False) #removing border lines

3.4 Change the color of bars

Matplotlib provides very comprehensive choices for you to choose colors. I am going to use plasma. You can try different color palette such as ‘viridis’, ‘plasma’, ‘inferno’, ‘magma’, ‘cividis’. If you would like to use one color with different shades, you can pass ‘Greys’, ‘Purples’, ‘Blues’, ‘Greens’, ‘Oranges’, ‘Reds’ to palette parameter.

fig,ax = plt.subplots(figsize=(9,6))
sns.barplot(x=’Province_Code’,y=’Minimum_Wage’,data=df,ci=95,ax=ax,
 order = df.sort_values(‘Minimum_Wage’).Province_Code, palette = ‘plasma’)
ax.set_title(‘Minimum Wage Comparison across Canada’)
ax.tick_params(labelsize=16,length=0)
plt.box(False)

3.5 Add grid lines

Adding grid lines will help audience to pinpoint the height of each bar and have a better understanding of the numbers.

fig,ax = plt.subplots(figsize=(9,6))
sns.barplot(x=’Province_Code’,y=’Minimum_Wage’,data=df,ci=95,ax=ax,
 order = df.sort_values(‘Minimum_Wage’).Province_Code, palette = ‘plasma’)
ax.set_title(‘Minimum Wage Comparison across Canada’)

ax.tick_params(labelsize=16,length=0)
plt.box(False)
# add grid lines for y axis
ax.yaxis.grid(linewidth=0.5,color=’black’)
# put the grid lines below bars
ax.set_axisbelow(True)

By default, the gridlines are solid lines. It can also be changed by pass variables to the parameter linestyle. You can check out more choices at the matplotlib page.

3.6 Beautify title, labels and ticks

In this step, I am going to beautify the title by giving it a background color and changing the font size and font color. Also, I am going to change the ticks and axis labels.

fig,ax = plt.subplots(figsize=(12,8))
sns.barplot(x=’Province_Code’,y=’Minimum_Wage’,data=df,ci=95,ax=ax,
 order = df.sort_values(‘Minimum_Wage’).Province_Code, palette = ‘plasma’,alpha=0.8)
ax.set_title(‘Minimum Wage Comparison across Canada’,backgroundcolor=’#565656',
 fontsize=20, weight=’bold’,color=’white’,style=’italic’,loc=’center’,pad=30)
 
ax.tick_params(labelsize=16,length=0)
plt.box(False)
ax.yaxis.grid(linewidth=0.5,color=’grey’,linestyle=’-.’)
ax.set_axisbelow(True)
ax.set_xlabel(‘Province’,weight=’bold’,size=15)
ax.set_ylabel(‘Minimum Wage’,weight=’bold’,size=15)

plt.yticks(np.arange(0, 18, 4)) #yticks starts from 0 and ends at 18, step is 4
ax.set_yticklabels([“ “,”$4",”$8",’$12',’$16'],color=’#565656')
plt.xticks(rotation=30,color=’#565656')
plt.show()

Now, you get a much more beautify bar chart.

At the very end, let us take a final look at the Before and After charts. The beautified chart is much more clear and and can help audience to read it more easily. Bars are in an ascending order. Labels and title are also informative. Grid lines is helpful in pinpointing the actual numbers of each bars.

If you are interested in making other types of beautiful charts, you may want to check out my other posts.