CODEWITHSUNDEEP
pythonseaborn
sjdh
sjdh

Reputation: 4006

Log-log lmplot with seaborn

Can Seaborn's lmplot plot on log-log scale?

This is lmplot with linear axes:

import numpy as np
import pandas as pd
import seaborn as sns
x =  10**arange(1, 10)
y = 10** arange(1,10)*2
df1 = pd.DataFrame( data=y, index=x )
df2 = pd.DataFrame(data = {'x': x, 'y': y}) 
sns.lmplot('x', 'y', df2)

sns.lmplot('x', 'y', df2):

sns.lmplot [2]:

Upvotes: 54

Views: 120499

Answers (4)

Rémy Hosseinkhan Boucher
Rémy Hosseinkhan Boucher

Reputation: 191

Generally, without explicit usage of matplotlib, use the subplot_kws argument in the facet_kws argument of the plot function.

From the seaborn documentation https://seaborn.pydata.org/generated/seaborn.lmplot.html

facet_kws: dict

Dictionary of keyword arguments for FacetGrid.

Example: facet_kws={"subplot_kws": {"yscale": "log"}} for a y-axis in log-scale. Of course, be sure data processed is positive.

Upvotes: 1

Alexander Braekevelt
Alexander Braekevelt

Reputation: 950

The simplest way to make a log-log plot from (probably) any seaborn plot is:

plt.xscale('log')
plt.yscale('log')

In the example:

import numpy as np
import pandas as pd
import seaborn as sns
x =  10**np.arange(1, 10)
y = 10** np.arange(1,10)*2
df1 = pd.DataFrame( data=y, index=x )
df2 = pd.DataFrame(data = {'x': x, 'y': y}) 
sns.lmplot('x', 'y', df2)
plt.xscale('log')
plt.yscale('log')

Link to resulting image

Upvotes: 13

mwaskom
mwaskom

Reputation: 49022

If you just want to plot a simple regression, it will be easier to use seaborn.regplot. This seems to work (although I'm not sure where the y axis minor grid goes)

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

x = 10 ** np.arange(1, 10)
y = x * 2
data = pd.DataFrame(data={'x': x, 'y': y})

f, ax = plt.subplots(figsize=(7, 7))
ax.set(xscale="log", yscale="log")
sns.regplot("x", "y", data, ax=ax, scatter_kws={"s": 100})

enter image description here

If you need to use lmplot for other purposes, this is what comes to mind, but I'm not sure what's happening with the x axis ticks. If someone has ideas and it's a bug in seaborn, I'm happy to fix it:

grid = sns.lmplot('x', 'y', data, size=7, truncate=True, scatter_kws={"s": 100})
grid.set(xscale="log", yscale="log")

enter image description here

Upvotes: 77

Paul H
Paul H

Reputation: 68186

Call the seaborn function first. It returns a FacetGrid object which has an axes attribute (a 2-d numpy array of matplotlib Axes). Grab the Axes object and pass that to the call to df1.plot.

import numpy as np
import pandas as pd
import seaborn as sns

x =  10**np.arange(1, 10)
y = 10**np.arange(1,10)*2
df1 = pd.DataFrame(data=y, index=x)
df2 = pd.DataFrame(data = {'x': x, 'y': y})

fgrid = sns.lmplot('x', 'y', df2)    
ax = fgrid.axes[0][0]
df1.plot(ax=ax)        

ax.set_xscale('log')
ax.set_yscale('log')

Upvotes: 6

Related Questions