Set Matplotlib colorbar size to match graph

python

matplotlib

colorbar

plotting

byAlex Kataev·Dec 13, 2024

Why worry about manually adjusting your Matplotlib colorbar when you can have it automatically size itself? Say hello to make_axes_locatable.

import matplotlib.pyplot as plt
from mpl_toolkits.axes_grid1 import make_axes_locatable

fig, ax = plt.subplots()
im = ax.imshow([[1, 2], [2, 3]])  # Put your data here

# Make colorbar as tall as the main plot.
divider = make_axes_locatable(ax)
cax = divider.append_axes("right", size="5%", pad=0.05)

# Add the colorbar to the created axes.
fig.colorbar(im, cax=cax)

plt.show()

This code ensures your colorbar scales in harmony with your graph. Now, isn't that pleasing to the eyes?

The secret sauce behind dynamic colorbar scaling

Half-baked {insert favorite food} doesn't taste good, right? Similarly, colorbars that don't match your graph's size can lead to unenthusiastic nods from your audience. To avoid that, we'll help you precisely control your colorbar sizing.

Chef's special: Ax Divider

make_axes_locatable makes sure your colorbar fits just right by adjusting its size and padding:

# We have defined `fig` and `ax` already

# Magic happens here. Adjust these to suit your figure size
divider = make_axes_locatable(ax)
cax = divider.append_axes("right", size="7%", pad=0.07)

# There you go. The cax is ready to serve … I mean plot
plt.colorbar(im, cax=cax)

Sous Chef's special: Axes via `add_axes`

ax.add_axes helps cut and serve the pie -- I mean, define axes -- to precise dimensions for the colorbar:

# Let's get the original graph axes position
pos = ax.get_position()

# Carefully prepare a take-out box for the colorbar
cax = fig.add_axes([pos.x1 + 0.01, pos.y0, 0.02, pos.height])

# In goes the colorbar, nice and comfy
plt.colorbar(im, cax=cax)

Don't forget to maintain aspect ratios!

Sharing axes can squeeze your graph like the last toothpaste drop. To avoid that, use separate axes for the colorbar to maintain the aspect ratio:

# Preserve the brownie-shaped graph while fitting the colorbar
fig.colorbar(im, cax=cax)

Automatic fraction and padding adjustments

fraction and pad in plt.colorbar are your sous-chefs in the kitchen. With some minor adjustments, you can make your code work for variable graph proportions:

fig.colorbar(im, fraction=0.046 if ax.get_aspect() == 'equal' else 0.035, pad=0.04)

Dressing up your plots: Advanced Tips

Confronting multiple subplots

Facing subplot overload? Fret not! The ax option helps maintain consistent colorbars across all your subplots, akin to maintaining a dress code at a party:

# `axs` is an array of subplots
fig.colorbar(im, ax=axs, fraction=0.046, pad=0.04)

Colorbar orientation: Horizontal vs Vertical

Your colorbar can be a horizontal limbo dancer or a vertical pole vaulter. Depending on its orientation, you might need to adjust your fraction and pad:

# For a horizontal colorbar at the bottom
cax = divider.append_axes("bottom", size="5%", pad=0.3)
plt.colorbar(im, cax=cax, orientation='horizontal')

Pro-tip: Reusable function

Frequently plotting? Roll out your colorbars in style with a reusable 'couture' function:

def whipUpColorbar(fig, ax, im, orientation='vertical'):
    divider = make_axes_locatable(ax)
    cax = divider.append_axes("right", size="5%", pad=0.05) if orientation=='vertical' else divider.append_axes("bottom", size="5%", pad=0.3)
    fig.colorbar(im, cax=cax, orientation=orientation)
# No chic party is complete without bespoke couture!