Adding units to heatmap annotation in Seaborn

Question

I am trying to show a table of percentages as a heatmap in Seaborn:

sns.heatmap(S, annot=True, fmt=".1f", linewidths=1.0, square=1)

However, I'd like the percent sign to appear after each number in the heatmap annotations. The fmt flag only seems to accept number format specifiers. Is there a way to do this either within Seaborn or with some matplotlib tweaking?

Serenity · Accepted Answer · 2016-06-03T02:38:49.157

38

You have to iterate over all text values of a heatmap and add % sign:

import matplotlib.pyplot as plt
import seaborn as sns
import numpy as np
from matplotlib.ticker import FuncFormatter

sns.set()
fig, ax0 = plt.subplots(1,1)
data = np.array([[10.01,20.20],[-0.23,0.],[23.1049,-100.000000]])
ax = sns.heatmap(data, annot=True, fmt = '.1f', square=1, linewidth=1.)
for t in ax.texts: t.set_text(t.get_text() + " %")
plt.show()

edited Jun 03 '16 at 02:38

answered Jun 02 '16 at 23:08

Serenity

35,289
20
120
115

I'm looking to add percent signs to the annotations (e.g. 0.23% in your figure), not to the colorbar. – Jsl Jun 02 '16 at 23:50
3

Converting '.1f' to '.1%' solved the problem for me. – Meghna Natraj Jul 08 '17 at 02:35
1

It is nice solution, but it will be better to change the label to percent as well. – Pengju Zhao Jul 28 '17 at 07:53
Do you know how to express in millions. e.g. 500k shows up as 0.5m – user147529 Feb 05 '21 at 11:11

score 27 · Answer 2 · answered Sep 27 '17 at 14:53

27

Replace the ".1f" to ".1%". And it should solve your problem.

sns.heatmap(S, annot=True, fmt=".1%", linewidths=1.0, square=1)

answered Sep 27 '17 at 14:53

Meghna Natraj

673
7
15

2

Then the colormap will be from 0 to 1 instead of 0 to 100. – ilija139 Apr 25 '19 at 02:26
Thanks! That's exactly what I was looking for on google! – Fernando Wittmann Apr 08 '21 at 19:36
@ilija139 You are right. Please see my answer that addresses the colorbar. p – Gobryas Feb 02 '23 at 20:25

score 1 · Answer 3 · answered Feb 02 '23 at 20:22

You will have to both adjust the format of your 'annotations' using the fmt parameter AND the 'colorbar' using cbar_kws.

from matplotlib.ticker import FuncFormatter
fmt = lambda x,pos:'{:.0%}'.format(x)
hm = sns.heatmap(data=S, annot=True, ax=ax1, fmt='.1%', cbar_kws={'format': FuncFormatter(fmt)})

For reference check this seaborn issue.

Adding units to heatmap annotation in Seaborn

3 Answers3

Linked