How to plot one linear regression scatterplot for multiple columns in a dataframe with seaborn?

Question

Lets say I have a dataframe df which has index 0 to 10, and 4 columns of data, c_1, c_2, c_3, and c_4. When I plot this as follows:

sns.regplot(data=df, x='c_1', y='c_2')
sns.regplot(data=df, x='c_3', y='c_4')

I get one graph on which there are two regression scatterplots. However, instead, I want one graph with only one regression scatterplot, keeping each indexed c_1 and c_3 as an x paired with c_2 and c_4 as y values, respectively. How can I do this? thanks for any help, let me know if I can make this more clear.

Mr. T · Accepted Answer · 2021-02-07 13:57:29Z

1

Most seaborn functions work better with long-form data. So, we could reorganize the data and label the categories that you imply with your code. Then, we plot a regplot for all data and plot a scatterplot on top for both categories:

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

#generate sample data
import numpy as np
np.random.seed(123)
n = 20
df = pd.DataFrame({"c_1":np.random.random(n), 
                   "c_2":np.zeros(n), 
                   "c_3":2 * np.random.random(n) - .5, 
                   "c_4":np.zeros(n)})
df.c_2 = 2 * df.c_1 - 3 * np.random.random(n)
df.c_4 = 3 * df.c_3 - 2 * np.random.random(n)        

#reformat the data for plotting...
df_plot = df[["c_1", "c_2"]].copy()
df_plot["cat"] = "set1" 
df_temp = df[["c_3", "c_4"]].copy()
df_temp.columns = ["c_1", "c_2"]
df_temp["cat"] = "set2" 
df_plot = df_plot.append(df_temp)

#...and plot
sns.regplot(data=df_plot, x='c_1', y='c_2', scatter=False)
sns.scatterplot(data=df_plot, x="c_1", y="c_2", hue="cat")

plt.show()

Sample output:

You may want to give the regplot a different color to avoid the impression the fitted line just applies to some of the data points in the scatter plot.

edited Feb 7, 2021 at 13:57

answered Feb 7, 2021 at 7:53

Mr. T

12.5k10 gold badges39 silver badges67 bronze badges

Sign up to request clarification or add additional context in comments.

2 Comments

mwaskom Over a year ago

You're getting a weird shadowing effect that I think will go away if you add scatter=False to the regplot call.

Mr. T Over a year ago

Excellent suggestion. When I first noticed it, I thought about increasing the marker size in the scatter plot (I did not know that you can turn the scatter plot off in regplot) but decided against it because, well, I like this effect. But obviously not everybody's cuppa, so I changed it.

Mohil Patel · Accepted Answer · 2021-02-07 08:09:52Z

1

You can use subplots for that.

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

plt.figure(figsize=(12,5))

#subplot(nrows, ncols, index)
plt.subplot(1, 2, 1)
sns.regplot(data=df, x='c_1', y='c_2')

plt.subplot(1, 2, 2)
sns.regplot(data=df, x='c_3', y='c_4') 

plt.show()

answered Feb 7, 2021 at 8:09

Mohil Patel

5203 silver badges9 bronze badges

Comments

ListenSoftware Louise Ai Agent · Accepted Answer · 2021-02-14 14:57:24Z

try a jointplot

 np.random.seed(123)
 n = 20
 df = pd.DataFrame({"c_1":np.random.random(n), 
               "c_2":np.zeros(n), 
               "c_3":2 * np.random.random(n) - .5, 
               "c_4":np.zeros(n)})
 df.c_2 = 2 * df.c_1 - 3 * np.random.random(n)
 df.c_4 = 3 * df.c_3 - 2 * np.random.random(n)        

 #reformat the data for plotting...
 df_plot = df[["c_1", "c_2"]].copy()
 df_plot["cat"] = "set1" 
 df_temp = df[["c_3", "c_4"]].copy()
 df_temp.columns = ["c_1", "c_2"]
 df_temp["cat"] = "set2" 
 df_plot = df_plot.append(df_temp)

 import scipy.stats as stats

 plt.title('combined sets')
 g=sns.JointGrid(data=df_plot, x='c_1', y='c_2')
 g.plot( sns.regplot,sns.scatterplot)
 g=g.plot_joint(sns.kdeplot)
 g=g.plot_marginals(sns.kdeplot,shade=True)
 plt.show()


 filter=df_plot['cat']=='set1'
 plt.title('Set1')
 g = (sns.jointplot(x="c_1",
         y="c_2",
         kind='scatter',
         color='green',
         data=df_plot[filter],
         marginal_kws=dict(bins=10, rug=True))
.plot_joint(sns.regplot)
.plot_joint(sns.kdeplot))
 plt.show()

filter=df_plot['cat']=='set2'
plt.title('Set2')
g = (sns.jointplot(x="c_1",
         y="c_2",
         kind='scatter',
         color='blue',
         data=df_plot[filter],
         marginal_kws=dict(bins=10, rug=True))
.plot_joint(sns.regplot)
.plot_joint(sns.kdeplot))

plt.show()

Collectives™ on Stack Overflow

How to plot one linear regression scatterplot for multiple columns in a dataframe with seaborn?

3 Answers 3

2 Comments

Comments

Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

3 Answers 3

2 Comments

Comments

Comments

Your Answer

Sign up or log in

Post as a guest

Related