The use case it to place a figure legend below a grid of subplots, while also using tight_layout() to adjust the spacing between the subplots. This is possible using this code example, where LEGEND_HEIGHT must be set to the fraction of figure height used by the figure:
import collections
import matplotlib.pyplot as plt
import numpy as np
LEGEND_HEIGHT = 0.05
# Create figure with a grid of plots and random data.
fig, axes = plt.subplots(5, 5, figsize=(10, 8))
for ax in axes.flatten():
ax.plot(np.cumsum(np.random.uniform(-1, 1, 100)), label='Line')
# Collect legend labels from all plots.
entries = collections.OrderedDict()
for ax in axes.flatten():
for handle, label in zip(*ax.get_legend_handles_labels()):
entries[label] = handle
# Adjust spacing between plots and make space for legend.
fig.tight_layout(rect=(0, LEGEND_HEIGHT, 1, 1), h_pad=0.5, w_pad=0.5)
# Add legend below the grid of plots.
legend = fig.legend(
entries.values(), entries.keys(),
loc='upper center', bbox_to_anchor=(0.5, LEGEND_HEIGHT))
fig.savefig('figure.png')
However, the number of subplots in the grid entries and entries in the legend frequently changes for me. Right now, this requires a lot of time for manually tuning LEGEND_HEIGHT to a suitable value. Does Matplotlib allow for an automatic solution here?
The use case it to place a figure legend below a grid of subplots, while also using
tight_layout()to adjust the spacing between the subplots. This is possible using this code example, whereLEGEND_HEIGHTmust be set to the fraction of figure height used by the figure:However, the number of subplots in the grid entries and entries in the legend frequently changes for me. Right now, this requires a lot of time for manually tuning
LEGEND_HEIGHTto a suitable value. Does Matplotlib allow for an automatic solution here?