A colorbar serves as a visual legend for scalar mappable objects in Matplotlib, translating color gradients into quantitative values. Configuring it properly is essential for accurate data interpretation in heatmaps, contour plots, and surface visualizations.

Instantiating a Basic Colorbar

The standard approach involves passing the mappablle artist returned by plotting functions direct to the figure's colorbar method.

import matplotlib.pyplot as plt
import numpy as np

# Generate sample 2D array
matrix = np.random.uniform(0, 100, size=(12, 12))
figure, main_axis = plt.subplots(figsize=(6, 5))
mesh_plot = main_axis.pcolormesh(matrix, cmap='plasma')

# Attach the legend to the figure
figure.colorbar(mesh_plot, ax=main_axis)
plt.show()

Here, pcolormesh generates the visual representation, and the resulting artist is linked to the colorbar renderer.

Positioning and Layout Control

By default, the bar aligns to the right of the parent axes. The location keyword argument shifts it to alternative boundaries such as 'bottom', 'top', or 'left'. Spatial relationships are further refined using pad (spacing from the main plot), shrink (length scaling), and aspect (ratio of long to short dimensions).

figure.colorbar(mesh_plot, ax=main_axis, location='bottom', pad=0.15, shrink=0.8, aspect=25)

Colormap Configuration and Discretization

While continuous gradients are standard, discrete color levels often improve readability for categorical or binned data. The get_cmap utility combined with the lut (lookup table) parameter restricts the gradient to a fixed number of steps.

from matplotlib.cm import get_cmap

# Restrict gradient to 6 distinct bands
stepped_cmap = get_cmap('cividis', lut=6)
bounded_plot = main_axis.imshow(matrix, cmap=stepped_cmap)
figure.colorbar(bounded_plot, ax=main_axis)

Annotating Ticks and Labels

Descriptive metadata clarifies the scale. The label paramter adds a longitudinal axis title, while explicit tick positioning overrides automatic interval selection. Custom string replacements for numeric ticks require accessing the colorbar's underlying axes object.

cbar = figure.colorbar(mesh_plot, ax=main_axis, ticks=[0, 25, 50, 75, 100])
cbar.set_label('Intensity Measurement', fontsize=12)
cbar.set_ticklabels(['Minimal', 'Low', 'Moderate', 'High', 'Peak'])

Note that tick label orientation automatically adjusts based on the colorbar's placement, but manual overrides can be applied via the underlying axes if necessary.

Isolating the Colorbar in a Dedicated Axes

Complex layouts often demand precise control over the colorbar's geometry. The axes_grid1 toolkit enables the creation of a partitioned subplot specifically for the legend, decoupling it from the automatic layout engine.

from mpl_toolkits.axes_grid1 import make_axes_locatable

figure, primary_ax = plt.subplots()
visual = primary_ax.contourf(np.random.randn(15, 15), cmap='inferno')

# Partition the existing axis
layout_splitter = make_axes_locatable(primary_ax)
dedicated_cbar_ax = layout_splitter.append_axes("right", size="4%", pad=0.25)

# Render the bar within the isolated axis
figure.colorbar(visual, cax=dedicated_cbar_ax)
plt.show()

This technique reserves exact fractional space for the legend, preventing distortion of the primary data visualization during window resizing or export.