Matplotlib make tick labels font size smaller

Question

In a matplotlib figure, how can I make the font size for the tick labels using ax1.set_xticklabels() smaller?

Further, how can one rotate it from horizontal to vertical?

Thanks for asking this question, I'm trying to solve this problem right now. I'm not strongly opinionated here, but looks like voters think the best answer is not the accepted one. What do you think? — Russia Must Remove Putin, Jun 18 '17 at 19:25
did I properly read that there are at least 5 different ways to change the fontsize in matplotlib? :D — famargar, Feb 26 '18 at 13:06

Autiwa · Answer 1 · 2019-10-07T13:16:12.873

703

There is a simpler way actually. I just found:

import matplotlib.pyplot as plt
# We prepare the plot  
fig, ax = plt.subplots()

# We change the fontsize of minor ticks label 
ax.tick_params(axis='both', which='major', labelsize=10)
ax.tick_params(axis='both', which='minor', labelsize=8)

This only answers to the size of label part of your question though.

edited Oct 07 '19 at 13:16

answered Jul 08 '12 at 19:37

Autiwa

7,172
2
11
10

25

Use `axis = 'x'` or `axis = 'y'` to modify just one of the axis – loved.by.Jesus Apr 28 '20 at 15:08

score 309 · Answer 2 · answered May 19 '17 at 19:58

309

To specify both font size and rotation at the same time, try this:

plt.xticks(fontsize=14, rotation=90)

answered May 19 '17 at 19:58

scottlittle

15,865
5
45
66

11

Note that this is used when working with MATLAB-like pyplot interface. In case of object-oriented interface tick_params() method of the axes object or set_tick_params() method of the axis object should be used (see other answers). – wombatonfire Jul 20 '18 at 17:45

score 157 · Accepted Answer · edited May 23 '17 at 10:31

157

Please note that newer versions of MPL have a shortcut for this task. An example is shown in the other answer to this question: https://stackoverflow.com/a/11386056/42346

The code below is for illustrative purposes and may not necessarily be optimized.

import matplotlib.pyplot as plt
import numpy as np

def xticklabels_example():
    fig = plt.figure() 

    x = np.arange(20)
    y1 = np.cos(x)
    y2 = (x**2)
    y3 = (x**3)
    yn = (y1,y2,y3)
    COLORS = ('b','g','k')

    for i,y in enumerate(yn):
        ax = fig.add_subplot(len(yn),1,i+1)

        ax.plot(x, y, ls='solid', color=COLORS[i]) 

        if i != len(yn) - 1:
            # all but last 
            ax.set_xticklabels( () )
        else:
            for tick in ax.xaxis.get_major_ticks():
                tick.label.set_fontsize(14) 
                # specify integer or one of preset strings, e.g.
                #tick.label.set_fontsize('x-small') 
                tick.label.set_rotation('vertical')

    fig.suptitle('Matplotlib xticklabels Example')
    plt.show()

if __name__ == '__main__':
    xticklabels_example()

enter image description here

edited May 23 '17 at 10:31

Community

1
1

answered Jun 17 '11 at 19:06

mechanical_meat

155,494
24
217
209

93

Just as an aside: Instead of looping through the tick label objects, you can use `plt.setp`. (Also, have a look at `ax.tick_params`) For example, you can just do `plt.setp(ax.get_xticklabels(), rotation='vertical', fontsize=14)`. Also, axes objects have an `ax.is_last_row()` method which can be handy in cases like your example. Instead of `if i != len...`, you can do `if not ax.is_last_row()`. (Why it's a method, I have no clue... Matplotlib hates properties, apparently!) – Joe Kington Jun 18 '11 at 02:11
1

@Joe: this is my first time hearing about `is_last_row()`, thank you! I've used `plt.setp` in the past, and wasn't necessarily sure it represented a more canonical way of customizing tick labels. You've removed that doubt, thanks. More broadly: your answers, especially in the `matplotlib` tag, frequently leave me in awe. Keep up the great work. – mechanical_meat Jun 18 '11 at 02:51
2

Thanks!! I appreciate the kind words, especially from you! Canonical is very relative, in this case. `plt.setp` is a "matlab-ism", and an explicit loop is probably much more pythonic. Being a matlab convert, myself, `setp` feels natural, but to each their own. Either one is quite readable, i.m.o. – Joe Kington Jun 18 '11 at 03:36
66

In new versions I found that you had to use the _labelsize_ parameter: e.g. `ax.tick_params(axis='x', labelsize=8)` – Vladtn Feb 16 '12 at 20:49
Should be `get_ticklabels` not `get_major_ticks` – jez Jun 15 '15 at 10:25

score 68 · Answer 4 · answered May 14 '15 at 19:18

68

Alternatively, you can just do:

import matplotlib as mpl
label_size = 8
mpl.rcParams['xtick.labelsize'] = label_size

answered May 14 '15 at 19:18

user2468932

901
6
6

Kyle Swanson · Answer 5 · 2019-12-06T17:14:45.937

Another alternative

I have two plots side by side and would like to adjust tick labels separately.

The above solutions were close however they were not working out for me. I found my solution from this matplotlib page.

ax.xaxis.set_tick_params(labelsize=20)

This did the trick and was straight to the point. For my use case, it was the plot on the right that needed to be adjusted. For the plot on the left since I was creating new tick labels I was able to adjust the font in the same process as seting the labels.

ie

ax1.set_xticklabels(ax1_x, fontsize=15)
ax1.set_yticklabels(ax1_y, fontsize=15)

thus I used for the right plot,

ax2.xaxis.set_tick_params(labelsize=24)
ax2.yaxis.set_tick_params(labelsize=24)

A minor subtlety... I know... but I hope this helps someone :)

Bonus points if anyone knows how to adjust the font size of the order of magnitude label.

Try this: `ax.yaxis.offsetText.set_fontsize(18)` – jayesh Oct 14 '21 at 22:10 — jayesh, Oct 14 '21 at 22:10

score 50 · Answer 6 · edited Jan 15 '19 at 09:06

50

plt.tick_params(axis='both', which='minor', labelsize=12)

edited Jan 15 '19 at 09:06

Georgy

9,972
7
57
66

answered Jun 09 '17 at 18:02

jingweimo

4,472
8
41
69

This is the only answer that works if you want to change the fontsize without setting new labels. – David Stansby Sep 20 '17 at 09:20
@DavidStansby I don't believe this is true. I had this exact issue( Mentioned above) and was able to use an alternate method different from this solution. using `ax.xaxis.set_tick_params(labelsize=20)` – Kyle Swanson Oct 15 '18 at 22:47
1

simplest answer, doesnt require reference to ax object – jeremy_rutman Jul 06 '19 at 08:54
I'd argue that you'd want ax object..but to each their own – Derek Eden Feb 19 '21 at 01:46

score 23 · Answer 7 · answered Jan 21 '16 at 09:20

23

In current versions of Matplotlib, you can do axis.set_xticklabels(labels, fontsize='small').

answered Jan 21 '16 at 09:20

Christoph

5,240
6
35
60

4

That forces you to specify the `labels` too. It's better if we can change just the font size. – Guimoute Nov 20 '19 at 13:47

score 16 · Answer 8 · edited Oct 01 '18 at 17:21

16

For smaller font, I use

ax1.set_xticklabels(xticklabels, fontsize=7)

and it works!

edited Oct 01 '18 at 17:21

Little Bobby Tables

4,126
4
26
45

answered Jun 16 '16 at 20:57

Pengyao

748
8
15

1

how do you define xticklabels? – mLstudent33 Feb 20 '20 at 19:21
@mLstudent33 it's just a list of labels..could be anything really – Derek Eden Feb 19 '21 at 01:46

score 15 · Answer 9 · answered Nov 19 '14 at 22:27

15

You can also change label display parameters like fontsize with a line like this:

zed = [tick.label.set_fontsize(14) for tick in ax.yaxis.get_major_ticks()]

answered Nov 19 '14 at 22:27

ATX_Bigfoot

159
1
3

2

I think that using list comprehensions that have side effects is not the best practice. I'd suggest using a standard for loop instead. – jjmontes Jul 31 '19 at 17:59

score 14 · Answer 10 · edited May 27 '18 at 07:46

14

The following worked for me:

ax2.xaxis.set_tick_params(labelsize=7)
ax2.yaxis.set_tick_params(labelsize=7)

The advantage of the above is you do not need to provide the array of labels and works with any data on the axes.

edited May 27 '18 at 07:46

Rohit Poudel

1,642
2
18
20

answered May 27 '18 at 04:20

Cryptoman

161
1
3

Matplotlib make tick labels font size smaller

10 Answers10

Linked

Related