pandas plot ylabel feature names doesn't fit

Question

I use pandas DataFrame to show feature importance plot.

However some features doesn't fit to ylabel due to their long names.

data = pd.DataFrame({'feature': feature,'importance': importance}).sort_values(by='importance', ascending=False)[0:10]
ax = data.plot(kind='barh', x='feature', y='importance')
ax.figure.savefig(filename)

score 1 · Answer 1 · answered Nov 10 '20 at 12:24

1

You can adjust the left parameter with matplotlib.pyplot.subplots_adjust:

from matplotlib import pyplot as plt

ax = data.plot(kind='barh', x='feature', y='importance')
plt.subplots_adjust(left=0.3)  # call after drawn the plot, but before showing or saving
ax.figure.savefig(filename)

answered Nov 10 '20 at 12:24

Niko Föhr

28,336
10
93
96

It worked but it crops from right side: https://prnt.sc/vgqg6s Is there any what to without crop? – tyasird Nov 10 '20 at 12:33
Did you try to also change the `right` parameter? Or, the figure size? – Niko Föhr Nov 10 '20 at 12:58

score 0 · Answer 2 · answered Nov 10 '20 at 12:44

Use plt.tight_layout() before saving, but after plotting, as follows:

import matplotlib.pyplot as plt

ax = data.plot(kind='barh', x='feature', y='importance')

plt.tight_layout()

ax.figure.savefig(filename)

More details in another thread: https://stackoverflow.com/a/17390833/13240220

tyasird · Accepted Answer · 2020-11-10T16:27:53.070

0

I found the solution: bbox_inches

ax.figure.savefig(filename, bbox_inches='tight')

edited Nov 10 '20 at 16:27

answered Nov 10 '20 at 13:33

tyasird

926
1
12
29

score 0 · Answer 4 · answered Nov 10 '20 at 15:39

0

Use constrained_layout; it is similar to tight_layout, but more flexible:

fig, ax = plt.subplots(constrained_layout=True)
data.plot(kind='barh', x='feature', y='importance', ax=ax)

answered Nov 10 '20 at 15:39

Jody Klymak

4,979
2
15
31

pandas plot ylabel feature names doesn't fit

4 Answers4