Remove the legend on a matplotlib figure

As of matplotlib v1.4.0rc4, a remove method has been added to the legend object.

Usage:

ax.get_legend().remove()

or

legend = ax.legend(...)
...
legend.remove()

See here for the commit where this was introduced.

If you want to plot a Pandas dataframe and want to remove the legend, add legend=None as parameter to the plot command.

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

df2 = pd.DataFrame(np.random.randn(10, 5))
df2.plot(legend=None)
plt.show()

You could use the legend's set_visible method:

ax.legend().set_visible(False)
draw()

This is based on a answer provided to me in response to a similar question I had some time ago here

(Thanks for that answer Jouni - I'm sorry I was unable to mark the question as answered... perhaps someone who has the authority can do so for me?)

if you call pyplot as plt

frameon=False is to remove the border around the legend

and '' is passing the information that no variable should be in the legend

import matplotlib.pyplot as plt
plt.legend('',frameon=False)

you have to add the following lines of code:

ax = gca()
ax.legend_ = None
draw()

gca() returns the current axes handle, and has that property legend_

If you are not using fig and ax plot objects you can do it like so:

import matplotlib.pyplot as plt

# do plot specifics
plt.legend('')
plt.show()

According to the information from @naitsirhc, I wanted to find the official API documentation. Here are my finding and some sample code.

I created a matplotlib.Axes object by seaborn.scatterplot(). The ax.get_legend() will return a matplotlib.legned.Legend instance. Finally, you call .remove() function to remove the legend from your plot.

ax = sns.scatterplot(......)
_lg = ax.get_legend()
_lg.remove()

If you check the matplotlib.legned.Legend API document, you won't see the .remove() function.

The reason is that the matplotlib.legned.Legend inherited the matplotlib.artist.Artist. Therefore, when you call ax.get_legend().remove() that basically call matplotlib.artist.Artist.remove().

In the end, you could even simplify the code into two lines.

ax = sns.scatterplot(......)
ax.get_legend().remove()

I made a legend by adding it to the figure, not to an axis (matplotlib 2.2.2). To remove it, I set the legends attribute of the figure to an empty list:

import matplotlib.pyplot as plt

fig = plt.figure()
ax1 = fig.add_subplot(111)
ax2 = ax1.twinx()

ax1.plot(range(10), range(10, 20), label='line 1')
ax2.plot(range(10), range(30, 20, -1), label='line 2')

fig.legend()

fig.legends = []

plt.show()

If you are using seaborn you can use the parameter legend. Even if you are ploting more than once in the same figure. Example with some df

import seaborn as sns

# Will display legend
ax1 = sns.lineplot(x='cars', y='miles', hue='brand', data=df)

# No legend displayed
ax2 = sns.lineplot(x='cars', y='miles', hue='brand', data=df, legend=None)

Here is a more complex example of legend removal and manipulation with matplotlib and seaborn dealing with subplots:

From seaborn, get the Axes object created by sns.<some_plot>() and do ax.get_legend().remove() as indicated by @naitsirhc. The following example also shows how to put the legend aside, and how to deal in a context of subplots.

# imports
import seaborn as sns
import matplotlib.pyplot as plt

# get data
sns.set()
sns.set_theme(style="darkgrid")
tips = sns.load_dataset("tips")

# subplots
fig, axes = plt.subplots(1, 2, sharex=True, sharey=True, figsize=(12,6)) 
fig.suptitle('Example of legend manipulations on subplots with seaborn')

g0 = sns.pointplot(ax=axes[0], data=tips, x="day", y="total_bill", hue="size")
g0.set(title="Pointplot with no legend")
g0.get_legend().remove() # <<< REMOVE LEGEND HERE 

g1 = sns.swarmplot(ax=axes[1], data=tips, x="day", y="total_bill", hue="size")
g1.set(title="Swarmplot with legend aside")
# change legend position: https://www.statology.org/seaborn-legend-position/
g1.legend(bbox_to_anchor=(1.02, 1), loc='upper left', borderaxespad=0)

https://i.stack.imgur.com/83bgv.png