JIYIK CN >

KDE is Kernel Density Estimateused to visualize the probability density of continuous and non-parametric data variables. When you want to visualize multiple distributions, KDEthe function will produce a more concise and easier to interpret plot.

Using KDE, we can visualize multiple data samples using a single chart, which is a more efficient way to visualize data.

SeabornSeaborn is a matplotlibpython library similar to . Seaborn can be integrated with pandasand numpyfor data representation.

Data scientists use this library to create informative and beautiful statistical charts and graphs. Using these presentations, you can understand the clear concepts and information flow within different modules.

We can plot univariate and bivariate plots using KDE function, Seaborn, and Pandas.

We will learn how to use pandas and seaborn for KDE plot visualization. This article will use mtcarsseveral samples of the dataset to demonstrate KDE plot visualization.

Before we get started, you need to install or add the seabornand sklearnlibraries using pip command.

pip install seaborn
pip install sklearn

Data Visualization Using Normal KDE Plot and Seaborn in Python

We can plot the data using normal KDE plotting functions with the Seaborn library.

In the example below, we create 1000 data samples using the random library and then arrange them in numpyan array of , as the Seaborn library is only available with numpyand Pandas dataframes.

Sample code:

import seaborn as sn
import matplotlib.pyplot as plt
import numpy as np

data = np.random.randn(1000)
# KDE Plot with seaborn
res = sn.kdeplot(data, color="red", shade="True")
plt.show()

Output:

We can also visualize the above data sample vertically or restore the above plot using KDE and Seaborn libraries. We use the plot attribute vertical=Trueto restore the above plot.

Sample code:

import seaborn as sn
import matplotlib.pyplot as plt
import numpy as np

data = np.random.randn(1000)
# KDE Plot with seaborn
res = sn.kdeplot(data, color="green", vertical=True, shade="True")
plt.show()

Output:

Plotting 1D KDE in Python using Pandas and Seaborn

We can use KDE plots to visualize the probability distribution of a single target or a continuous attribute. In the following example, we read mtcarsthe CSV file of the dataset.

There are over 350 entries in our dataset and we will visualize the univariate distribution along the x-axis.

Sample code:

import seaborn as sn
import matplotlib.pyplot as plt
import numpy as np
import pandas as pd

# read CSV file of dataset using pandas
dataset = pd.read_csv(r"C:\\Users\\DELL\\OneDrive\\Desktop\\samplecardataset.csv")
# kde plot using seaborn
sn.kdeplot(data=dataset, x="hp", shade=True, color="red")
plt.show()

Output:

You can also flip the plot by visualizing the data variable along the y-axis.

Sample code:

import seaborn as sn
import matplotlib.pyplot as plt
import numpy as np
import pandas as pd

# Read CSV file using pandas
dataset = pd.read_csv(r"C:\\Users\\DELL\\OneDrive\\Desktop\\samplecardataset.csv")
# KDE plotting using seaborn
sn.kdeplot(data=dataset, y="hp", shade=True, color="red")
plt.show()

Output:

We can visualize the probability distribution of multiple target values ​​in a single plot.

Sample code:

import seaborn as sn
import matplotlib.pyplot as plt
import numpy as np
import pandas as pd

# Read CSV file using pandas
dataset = pd.read_csv(r"C:\\Users\\DELL\\OneDrive\\Desktop\\samplecardataset.csv")
# KDE plotting using seaborn
sn.kdeplot(data=dataset, x="hp", shade=True, color="red")
sn.kdeplot(data=dataset, x="mpg", shade=True, color="green")
sn.kdeplot(data=dataset, x="disp", shade=True, color="blue")
plt.show()

Output:

Plotting 2D or Bivariate KDE Plots in Python Using Pandas and Seaborn

We can visualize the data in a two-dimensional or bivariate KDE plot using seaborn and pandas libraries.

In this way, we can visualize the probability distribution of a given sample for multiple continuous attributes. We visualize the data along the x and y axes.

Sample code:

import seaborn as sn
import matplotlib.pyplot as plt
import numpy as np
import pandas as pd

# Read CSV file using pandas
dataset = pd.read_csv(r"C:\\Users\\DELL\\OneDrive\\Desktop\\samplecardataset.csv")
# KDE plotting using seaborn
sn.kdeplot(data=dataset, shade=True, x="hp", y="mpg")
plt.show()

Output:

Similarly, we can plot the probability distribution of multiple samples using a single KDE plot.

Sample code:

import seaborn as sn
import matplotlib.pyplot as plt
import numpy as np
import pandas as pd

# Read CSV file using pandas
dataset = pd.read_csv(r"C:\\Users\\DELL\\OneDrive\\Desktop\\samplecardataset.csv")
# KDE plotting using seaborn
sn.kdeplot(data=dataset, shade=True, x="hp", y="mpg", cmap="Blues")
sn.kdeplot(data=dataset, shade=True, x="hp", y="cyl", cmap="Greens")
plt.show()

Output:

in conclusion

We have demonstrated KDE plot visualization using Pandas and Seaborn libraries in this tutorial. We have seen how to visualize probability distributions of single and multiple samples in a 1D KDE plot.

We discussed how to visualize two-dimensional data using KDE plots with Seaborn and Pandas.