Press "Enter" to skip to content

How to Make a Scatter Plot in Python using Seaborn

Data visualization is a big part of the process of data analysis. In this post, we will learn how make a scatter plot using Python and the package Seaborn. In detail, we will learn how to use the Seaborn methods scatterplot, regplot, lmplot, and pairplot to create scatter plots in Python.

More specifically, we will learn how to make scatter plots, change the size of the dots, change the markers, the colors, and change the number of ticks.  Furthermore, we will learn how to plot a regression line, add text, plot a distribution on a scatter plot, among other things. Finally, we will also learn how to save Seaborn plots in high resolution. That is, we learn how to make print-ready plots.

Scatter plots are powerful data visualization tools that can reveal a lot of information. Thus, this Python scatter plot tutorial will start explain what they are and when to use them. After we done that, we will learn how to make scatter plots.

What is a Scatter Plot?

A scatter plot is a two-dimensional (bivariate) data visualization that uses dots to represent the values gathered for two different variables. That is, one of the variables is plotted along the x-axis and the other plotted along the y-axis. For example, this scatter plot shows the relationship between a child’s height and the parents height.

Note, there are of course possible to create a scatter plot with other programming languages, or applications. In this post, we focus on how to create a scatter plot in Python but the user of R statistical programming language can have a look at the how to make a scatter plot in R tutorial.

Python Scatter Plot Tutorial

This is exactly what we are going to learn in this tutorial; how to make a scatter plot using Python and Seaborn.

In the first code chunk, below, we are importing the packages we are going to use. Furthermore, to get data to visualize (i.e., create scatter plots) we load this from an URL. Note, the %matplotlib inline code is only needed if we want to display the scatter plots in a Jupyter Notebook. This is also the case for the import warnings and warnings.filterwarnings(‘ignore’) part of the code.

%matplotlib inline
import numpy as np
import pandas as pd
import seaborn as sns

# Suppress warnings
import warnings

# Import Data
data = ''
df = pd.read_csv(data, index_col=0)


In the example, above, we used Pandas read_csv method to load data into a dataframe. If you want to find out more you can see the Pandas Read CSV Tutorial. If you want to load the data from the the web (e.g., parse HTML tables or JSON files):

How to use Seaborn to create Scatter Plots in Python

In this section we will learn how to make scatter plots using Seaborns scatterplot, regplot, lmplot, and pairplot methods.

Seaborn Scatter plot using the scatterplot method

First, we start with the most obvious method to create scatter plots using Seaborn: using the scatterplot method. In the first Seaborn scatter plot example, below, we plot the variables wt (x-axis) and mpg (y-axis). This will give us a simple scatter plot:

sns.scatterplot(x='wt', y='mpg', data=df)

How to Change the Size of the Dots

If we want to have the size of the dots represent one of the variables this is possible.  So, how do you change the size of the dots in a Seaborn plot? In the next example, we change the size of the dots using the size argument.

sns.scatterplot(x='wt', y='mpg',
                size='wt', data=df)
Scatter plot with larger dots

How to Change the Axis

In the next Seaborn scatter plot example we are going to learn how to change the ticks on the x- and y-axis’. That is, we are going to change the number of ticks on each axis. To do this we use the set methods and use the arguments xticks and yticks:

ax = sns.scatterplot(x='wt', y='mpg', size='wt', data=df)
ax.set(xticks=np.arange(1, 6, 1),
      yticks=np.arange(5, 50, 10))
Scatter diagram with customized axis’

Grouped Scatter Plot in Seaborn

If we have a categorical variable (i.e., a factor) and want to group the dots in the scatter plot we use the hue argument:

sns.scatterplot(x='wt', y='mpg',
                hue='vs', data=df)
Grouped scatter plot in Python

There is, of course, possible to group the data using Pandas groupby method. If you are interested in learning more about working grouping data using Pandas:

Changing the Markers (the dots)

Now, one way to change the look of the markers is to use the style argument.

data = ''
df = pd.read_csv(data, index_col=0)
sns.scatterplot(x='IQbio', y='IQfoster', style='class',
                hue='class', data=df)
Scattergraph with changed markers

Note, should also be possible to use the markers argument. However, it seems not to work, right now.

Bivariate Distribution on Scatter plot

In the next Seaborn scatter plot example we are going to plot a bivariate distribution on top of our scatter plot. This is quite simple and after we have called the scatterplot method we use the kdeplot method:

data = ''
df = pd.read_csv(data, index_col=0)

sns.scatterplot(x='wt', y='mpg', data=df)
sns.kdeplot(df.wt, df.mpg)
Bivariate distribution on a scatter plot

Seaborn Scatter plot using the regplot method

If we want a regression line (trend line) plotted on our scatter plot we can also use the Seaborn method regplot. In the first example, using regplot, we are creating a scatter plot with a regression line. Here, we also get the 95% confidence interval:

sns.regplot(x='wt', y='mpg',
            ci=None, data=df)
Regression line on scatter plot

Scatter Plot Without Confidence Interval

We can also choose to create a scatter plot without the confidence interval. This is simple, we just add the ci argument and set it to None:

sns.regplot(x='wt', y='mpg',
            fit_reg=False, data=df)
Without confidence interval

nterval we just change the ci=None to ci=70.

Seaborn regplot Without Regression Line

Furthermore, it’s possible to create a scatter plot without the regression line using the regplot method. In the next example, we just add the argument reg_fit and set it to False:

ax = sns.regplot(x='wt', y='mpg', data=df)
ax.set(xticks=np.arange(1, 6, 1),
      yticks=np.arange(5, 50, 10))

Changing the Number of Ticks

It’s also possible to change the number of ticks when working with Seaborn regplot. In fact, it’s as simple as working with the scatterplot method, we used earlier. To change the ticks we use the set method and the xticks and yticks arguments:

ax = sns.regplot(x='wt', y='mpg', data=df)
ax.set(xticks=np.arange(1, 6, 1),
      yticks=np.arange(5, 50, 10))

Changing the Color on the Scatterplot

If we want to change the color of the plot, it’s also possible. We just have to use the color argument. In the regplot example below we are getting a green trend line and green dots:

sns.regplot(x='wt', y='mpg',
            color='g', data=df) 

Adding Text to a Seaborn Plot

In the next example, we are going to learn how to add text to a Seaborn plot. Remember the plot, in the beginning, showing the relationship between children’s height and their parents height? In that plot there was a regression line as well as the correlation coeffienct and p-value. Here we are going to learn how to add this to a scatter plot. First, we start by using pearsonr to calculate the correlation between the variables wt and mpg:

from scipy.stats import pearsonr

corr = pearsonr(df['wt'], df['mpg'])
corr = [np.round(c, 2) for c in corr]
# Output: [-0.87, 0.0]

In the next code chunk we are creating a string variable getting our pearson r and p-value (corr[0] and corr[1], respectively). Next, we create the scatter plot with the trend line and, finally, add the text. Note, the first tow arguments are the coordinates where we want to put the text:

text = 'r=%s, p=%s' % (corr[0], corr[1])
ax = sns.regplot(x="wt", y="mpg", data=df)
ax.text(4, 30, text, fontsize=12)
Scatter plot with text (correlation)

Seaborn Scatter Plot with Trend Line and Density

It’s, of course, also possible to plot a bivariate distrubiton on the scatter plot with a regression line. In the final example, using Seaborn regplot we just add the kernel density estimation using Seaborn kde method:

import seaborn as sns

sns.regplot(x='wt', y='mpg',
            ci=None, data=df)
sns.kdeplot(df.wt, df.mpg)

How to Rotate the Axis Labels on a Seaborn Plot

If the labels on the axis’ are long, we may want to rotate them for readability. It’s quite simple to rotate the axis labels on the regplot object. Here we rotatelabels on the x-axis 90 degrees:

ax = sns.regplot(x='wt', y='mpg', data=df)
for item in ax.get_xticklabels():
Rotated ticks labels on Seaborn plot

Scatter Plot Using Seaborn lmplot Method

In this section we are going to cover Seaborn’s lmplot method. This method enables us create grouped Scatter plots with a regression line.

In the first example, we are just creating a Scatter plot with the trend lines for each group. Note, here we use the argument hue to set what variable we want to group by.

sns.lmplot(x='wt', y='mpg',
           hue='vs', data=df)

Changing the Color Using the Palette Argument

When working with Seaborn’s lmplot method we can also change the color of the lines, CIs, and dots using the palette argument. See here for different palettes to use.

sns.lmplot(x='wt', y='mpg', hue='vs',
           palette="Set1", data=df)

Change the Markers on the lmplot

As previously mentioned, when creating scatter plots, the markers can be changed using the markers argument. In the next Seaborn lmplot example we are changing the markers to crosses:

sns.lmplot(x='wt', y='mpg',
           hue='vs', palette="Set1", markers='+', data=df)

Furthermore, it’s also possible to change the markers of two, or more, categories. In the example below we use a list with the cross and “star” and get a plot with crosses and stars, for each group.

sns.lmplot(x='wt', y='mpg',
           palette="Set1", markers=['+', '*'], data=df)

See here for different markers that can be used with Seaborn plots.

Seaborn Pairplot: Scatterplot + Histogram

In the last section, before learning how to save high resolution images, we are going to use Seaborn’s pairplot method. The pairplot method will create pairs of plots

data = ''
df = pd.read_csv(data, index_col=0)
df = dfdf[['mpg', 'wt', 'hp', 'qsec']
Pairplot with Grouped Data

If we want to group the data, with different colors for different categories, we can use the argument hue, also with pairplot. In this Python data visualization example we also use the argument vars with a list to select which variables we want to visualize:

cols = ['mpg', 'wt', 'hp', 'qsec']
sns.pairplot(df, vars=cols, hue='vs')

Saving a High Resolution Plot in Python

If we are planning to publish anything we need to save the data visualization(s) to a high resolution file. In the last code example, below, we will learn how to save a high resolution image using Python and matplotlib.

Basically, there are three steps for creating and saving a Seaborn plot:

  1. Load the data using Pandas (e.g., using read_csv)
  2. Create the plot using Seaborn (e.g., using sns.pairplot)
  3. Use matplotlib (plt.savefig) to save the figure

Here’s a example of how to save a high resolution figure:

ax = sns.pairplot(df, vars=cols, hue='vs')
plt.savefig('pairplot.eps', format='eps', dpi=300)

Here’s a link to a Jupyter Notebook containing all of the above code.


In this post we have larned how to make scatter plots in Seaborn. Moreover, we have also learned how to:

  • work with the Seaborn’s scatterplot, regplot, lmplot, and pairplot methods
  • change color, number of axis ticks, the markers, and how to rotate the axis labels of Seaborn plots
  • save a high resolution, and print ready, image of a Seaborn plot

Be First to Comment

Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.

%d bloggers like this: