Plotting with Matplotlib and Seaborn

Learning Objectives

• Data visualization with seaborn and matplotlib
• Chapter 4 of Python Data Science Handbook.

Python Overview

In R I Want	In Python I Use
Base R	numpy
dplyr/tidyr	pandas
ggplot2	matplotlib/seaborn

Import Matplotlib and Seaborn, and Load Dataset

R

library(ggplot2)
data("mpg")

All other code will be Python unless otherwise marked.

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

Show and clear plots.

• Use plt.show() to display a plot.

• Use plt.clf() to clear a figure when making a new plot.

One Quantitative Variable: Histogram

• sns.histplot() makes a histogram.

  sns.histplot(x='hwy', data=mpg)
  plt.show()

  

  plt.clf()

  

One Categorical Variable: Barplot

• Use sns.countplot() to make a barplot to look at the distribution of a categorical variable:

  sns.countplot(x='class', data=mpg)
  plt.show()

  

  plt.clf()

  

One Quantitative Variable, One Categorical Variable: Boxplot

• Use sns.boxplot() to make boxplots:

  sns.boxplot(x='class', y='hwy', data=mpg)
  plt.show()

  

  plt.clf()

  

Two Quantitative Variables: Scatterplot

• Use sns.scatterplot() to make a basic scatterplot.

  sns.scatterplot(x='displ', y='hwy', data=mpg)
  plt.show()

  

  plt.clf()

  

Lines/Smoothers

• Use sns.regplot() to make a scatterplot with a regression line or a loess smoother.

• Regression line with 95% Confidence interval

  sns.regplot(x='displ', y='hwy', data=mpg)
  plt.show()

  

  plt.clf()

  

• Loess smoother with confidence interval removed.

  sns.regplot(x='displ', y='hwy', data=mpg, lowess=True, ci='None')
  plt.show()

  

  plt.clf()

  

Annotating by Third Variable

• Use the hue or style arguments to annotate by a categorical variable:

  sns.scatterplot(x='displ', y='hwy', hue='class', data=mpg)
  plt.show()

  

  plt.clf()

  

  sns.scatterplot(x='displ', y='hwy', style='class', data=mpg)
  plt.show()

  

  plt.clf()

  

• Use the hue or size arguments to annotate by a quantitative variable:

  sns.scatterplot(x='cty', y='hwy', hue='displ', data=mpg)
  plt.show()

  

  plt.clf()

  

  sns.scatterplot(x='cty', y='hwy', size='displ', data=mpg)
  plt.show()

  

  plt.clf()

  

Two Categorical Variables: Mosaic Plot

• Usually, you should just show a table of proportions when you have two categorical variables.

  pd.crosstab(mpg['class'], mpg['drv'], normalize='all')

  ## drv                4         f         r
  ## class                                   
  ## 2seater     0.000000  0.000000  0.021368
  ## compact     0.051282  0.149573  0.000000
  ## midsize     0.012821  0.162393  0.000000
  ## minivan     0.000000  0.047009  0.000000
  ## pickup      0.141026  0.000000  0.000000
  ## subcompact  0.017094  0.094017  0.038462
  ## suv         0.217949  0.000000  0.047009

  pd.crosstab(mpg['class'], mpg['drv'], normalize='index')

  ## drv                4         f         r
  ## class                                   
  ## 2seater     0.000000  0.000000  1.000000
  ## compact     0.255319  0.744681  0.000000
  ## midsize     0.073171  0.926829  0.000000
  ## minivan     0.000000  1.000000  0.000000
  ## pickup      1.000000  0.000000  0.000000
  ## subcompact  0.114286  0.628571  0.257143
  ## suv         0.822581  0.000000  0.177419

  pd.crosstab(mpg['class'], mpg['drv'], normalize='columns')

  ## drv                4         f     r
  ## class                               
  ## 2seater     0.000000  0.000000  0.20
  ## compact     0.116505  0.330189  0.00
  ## midsize     0.029126  0.358491  0.00
  ## minivan     0.000000  0.103774  0.00
  ## pickup      0.320388  0.000000  0.00
  ## subcompact  0.038835  0.207547  0.36
  ## suv         0.495146  0.000000  0.44

Facets

• Use sns.FacetGrid() followed by the map_dataframe() method to plot facets. You pass arguments to the plot (sns.histplot() or sns.scatterplot() etc) inside the map function.

  g = sns.FacetGrid(data=mpg, row='drv', col='class')
  g = g.map_dataframe(sns.histplot, x = 'hwy', kde=False)
  plt.show()

  

  plt.clf()

  

Labels

• Assign plot to an object. Then use the set_*() methods to add labels.

  scatter = sns.scatterplot(x='displ', y='hwy', data=mpg)
  scatter.set_xlabel('Displacement')
  scatter.set_ylabel('Highway')
  scatter.set_title('Highway versus Displacement')
  plt.show()

  

Saving Figures

1. First, assign a figure to an object.

   scatter = sns.scatterplot(x='displ', y='hwy', data=mpg)

2. Extract the figure. Assign this to an object.

   fig = scatter.get_figure()

3. Save the figure.

   fig.savefig('./scatter.pdf')

• You can do all of these steps using piping.

  sns.scatterplot(x='displ', y='hwy', data=mpg) \
    .get_figure() \
    .savefig('./scatter.pdf')