5 Best Ways to Create a Scatter Plot with Seaborn, Python Pandas

by

💡 Problem Formulation: When working with datasets in Python, data visualization becomes a vital step for understanding trends and patterns. Creating a scatter plot is a fundamental technique for exploring the relationship between two numerical variables. This article outlines five methods to create a scatter plot using the Seaborn library, which works harmoniously with Pandas DataFrames. For a dataset with columns ‘age’ and ‘income’, the desired output is a graphical representation elucidating the correlation between these variables.

Method 1: The Classic Scatter Plot

This method involves using Seaborn’s built-in scatterplot() function, which is a versatile and straightforward way to generate scatter plots. It takes at least two arguments: the x and y axes data points which can be columns from a Pandas DataFrame. You can customize the aesthetics and behavior of the plot with additional parameters.

Here’s an example:

import seaborn as sns
import pandas as pd
import matplotlib.pyplot as plt

# Sample data
data = pd.DataFrame({
    'age': [25, 30, 35, 40, 45],
    'income': [50000, 55000, 60000, 65000, 70000]
})

# Creating the scatter plot
sns.scatterplot(x='age', y='income', data=data)
plt.show()

The output is a window displaying a scatter plot with ‘age’ on the x-axis and ‘income’ on the y-axis.

This simple example uses a small dataset and creates a scatter plot showing the relationship between age and income. Using Seaborn’s scatterplot() function, it’s easy to see at a glance how these variables relate on a 2D graph, with many additional options available to enhance the visualization.

Method 2: Scatter Plot with Categories

Seaborn allows plotting categorial variables against numerical data by color-coding the points. The hue parameter in the scatterplot() function can differentiate data points based on a categorical variable, providing depth to the scatter plot.

Here’s an example:

import seaborn as sns
import pandas as pd
import matplotlib.pyplot as plt

# Sample data
data = pd.DataFrame({
    'age': [25, 30, 35, 40, 45, 25, 30, 35, 40, 45],
    'income': [50000, 54000, 61000, 68000, 72000, 52000, 58000, 65000, 63000, 75000],
    'gender': ['Male', 'Male', 'Male', 'Male', 'Male', 'Female', 'Female', 'Female', 'Female', 'Female']
})

# Creating the scatter plot with categories
sns.scatterplot(x='age', y='income', hue='gender', data=data)
plt.show()

The output is a scatter plot with ‘age’ on the x-axis and ‘income’ on the y-axis, with points colored by the ‘gender’ category.

This code snippet extends the basic scatter plot by categorizing data points using the ‘gender’ column. By specifying the hue parameter, we distinguish between male and female data points, enabling a comparison of income distribution across ages for both genders.

Method 3: Scatter Plot with Regression Line

For a scatter plot that includes a regression line, indicating the trend, we can use Seaborn’s regplot(). This method will automatically calculate and plot a regression line, which is useful for highlighting relationships in your data.

Here’s an example:

import seaborn as sns
import pandas as pd
import matplotlib.pyplot as plt

# Sample data
data = pd.DataFrame({
    'age': [20, 25, 30, 35, 40, 45, 50],
    'income': [40000, 45000, 50000, 55000, 60000, 65000, 70000]
})

# Scatter plot with a regression line
sns.regplot(x='age', y='income', data=data)
plt.show()

The output is a scatter plot with ‘age’ on the x-axis and ‘income’ on the y-axis, complemented by a regression line through the data points.

This code snippet creates a scatter plot while simultaneously fitting and displaying a regression line, hence conveying both the data distribution and the trend. The regplot() function from Seaborn makes it straightforward to carry out both tasks in one step.

Method 4: Scatter Plot with Multiple Variables

Seaborn provides the capability to explore multidimensional relationships using scatter plots by employing multiple aesthetic changes. The scatterplot() function can modify the size (size) and style (style) of the data points in addition to the color (hue).

Here’s an example:

import seaborn as sns
import pandas as pd
import matplotlib.pyplot as plt

# Sample data with an additional 'savings' variable
data = pd.DataFrame({
    'age': [25, 30, 35, 40, 45],
    'income': [50000, 55000, 60000, 65000, 70000],
    'savings': [20000, 25000, 30000, 35000, 40000]
})

# Creating the scatter plot with multiple variables
sns.scatterplot(x='age', y='income', size='savings', data=data, legend='brief')
plt.show()

The output is a scatter plot with varying point sizes representing different ‘savings’ amounts.

In this example, the size of each data point corresponds to the ‘savings’ value, creating a three-dimensional effect on a 2D plane. This scatter plot communicates more information about the dataset by combining three variables in a single visualization.

Bonus One-Liner Method 5: Scatter Plot Matrix

For an all-encompassing view that displays scatter plots for each pair of variables in the dataset, Seaborn’s pairplot() comes in handy. It’s particularly useful for a high-level overview of potential correlations in multidimensional data.

Here’s an example:

import seaborn as sns
import pandas as pd

# Sample data
data = pd.DataFrame({
    'age': [24, 29, 34, 39, 44],
    'income': [48000, 53000, 58000, 63000, 68000],
    'savings': [18000, 23000, 28000, 33000, 38000]
})

# Scatter plot matrix
sns.pairplot(data)
plt.show()

The output is a matrix of scatter plots for each pair of ‘age’, ‘income’, and ‘savings’ variables, along with histograms for individual variables on the diagonal.

This one-liner code quickly generates a comprehensive scatter plot matrix which provides a visual summary of linear relationships and variable distributions within a dataset. The pairplot() function is an efficient way to assess multiple relationships at once.

Summary/Discussion

  • Method 1: Classic Scatter Plot. Simple, best for initial data exploration. Lack of advanced features.
  • Method 2: Categorical Scatter Plot. Adds depth by color-coding categories. Limited to categorical differentiation.
  • Method 3: Scatter Plot with Regression Line. Provides trend insight with a fitted line. Assumes linear relationship; may not be suitable for all datasets.
  • Method 4: Multiple Variables Scatter Plot. Conveys complex, multi-dimensional data. Can become cluttered with too many variables.
  • Method 5: Scatter Plot Matrix. Quick, high-level overview of pair-wise relationships. Overwhelming for large datasets with many variables.