CODEWITHSUNDEEP
pythonseaborn
xxxxx
xxxxx

Reputation: 11

sns pairplot variable 'GrLivArea' output is blank

I got my housing data from kaggle. I am trying to do a pairplot using seaborn. However I am not sure why GrLivArea graph output is blank.

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

df = pd.read_csv('train.csv')

sns.pairplot(df, x_vars=['GrLivArea', 'LotArea', 'TotalBsmtSF', 'GarageArea'], y_vars=['SalePrice'], height=8, aspect = 1, kind='reg')
plt.show()

Able to advise?

Upvotes: 1

Views: 621

Answers (2)

ginagigo123
ginagigo123

Reputation: 31

I encounter the same problem when using sns.pairplot.

Then I find the solution on Github here.

By adding the parameter diag_kind=None in the sns.pairplot then the first plot could successfully build.

Upvotes: 2

Naman Doctor
Naman Doctor

Reputation: 182

This is because your x-axis for the first plot starts from 0.0 and goes till 1.0. If you do the following:

sns.pairplot(df, x_vars=['LotArea', 'TotalBsmtSF', 'GarageArea','GrLivArea'], y_vars=['SalePrice'], height=8, aspect = 1, kind='reg')

then the plot will be empty for LotArea.

At first, I assumed it was due to empty values (NaNs) in your dataset. But, I checked it, and it does not seem to be the case.

The following solution should solve your problem:

fig, axs = plt.subplots(ncols=4)
sns.regplot(x='GrLivArea', y='SalePrice', data=df, ax=axs[0])
sns.regplot(x='LotArea', y='SalePrice', data=df, ax=axs[1])
sns.regplot(x='TotalBsmtSF',y='SalePrice', data=df, ax=axs[2])
sns.regplot(x='GarageArea', y='SalePrice', data=df, ax=axs[3])
plt.show()

While it is not a pairplot, you can plot each regplot as a subplot.

Upvotes: 1

Related Questions