Adding caption below X-axis for a scatter plot using matplotlib

Question

I am pretty new to python and to the matplotlib library. I have created a scatter plot using matplotlib and now I wish to add caption a little below the X-axis. This is my code:

from matplotlib import pyplot as plt
import numpy as np
from pylab import *

file = open('distribution.txt', 'r')

txt="I need the caption to be present a little below X-axis"

x=[]
y=[]
for line in file:
    new=line.rstrip()
    mystring=new.split("\t")
    x.append(mystring[0])
    y.append(mystring[1])


fig = plt.figure()
ax1 = fig.add_axes((0.1,0.4,0.8,0.5))
ax1.set_title("This is my title")
ax1.set_xlabel('X-axis')
ax1.set_ylabel('Y-axis')
ax1.scatter(x,y, c='r')
fig.text(.05,.05,txt)
plt.xlim(0, 1.05)
plt.ylim(0, 2.5)
plt.show()

As you can see in the image my caption is way below the scatter plot, is there a way to bring it exactly below the X-axis? Also my scatter plot looks rectangular, is there a way to make it square like?

Answer 1

You can simply use figtext. You can also change the value of x and y-axes as you want.

txt="I need the caption to be present a little below X-axis"
plt.figtext(0.5, 0.01, txt, wrap=True, horizontalalignment='center', fontsize=12)

Answer 2

Something like:

from matplotlib import pyplot as plt
import numpy as np

txt="I need the caption to be present a little below X-axis"

# make some synthetic data
x = np.linspace(0, 1, 512)
y = np.random.rand(512)*2.3 + .1

fig = plt.figure()
ax1 = fig.add_axes((0.1, 0.2, 0.8, 0.7))

ax1.set_title("This is my title")
ax1.set_xlabel('X-axis')
ax1.set_ylabel('Y-axis')

# make the edge colors match the facecolors
ax1.scatter(x,y, c='r', edgecolors='face')
# center text
fig.text(.5, .05, txt, ha='center')

# use OO interface    
ax1.set_xlim([0, 1.05])
ax1.set_ylim([0, 2.5])

# resize the figure to match the aspect ratio of the Axes    
fig.set_size_inches(7, 8, forward=True)

plt.show()

might work. Making this easier to do is on the radar for mpl upstream, but we are still looking for someone to do it.

Answer 3

First, I feel weird posting an answer against the co-lead developer of matplotlib. Obviously, @tacaswell knows matplotlib far better than I ever will. But at the same time, his answer wasn't dynamic enough for me. I needed a caption that would always be based on the position of the xlabel, and couldn't just use text annotations.

I considered simply changing the xlabel to add a newline and the caption text, but that wouldn't clearly differentiate the caption, and you can't do things like change the text size or make it italic in the middle of a text string.

I solved this by using matplotlib's TeX capabilities. Here's my solution:

from matplotlib import pyplot as plt
from matplotlib import rc
import numpy as np
from pylab import *

rc('text', usetex=True)

file = open('distribution.txt', 'r')

txt="I need the caption to be present a little below X-axis"

x=[]
y=[]
for line in file:
    new=line.rstrip()
    mystring=new.split("\t")
    x.append(mystring[0])
    y.append(mystring[1])


fig = plt.figure()
ax1 = fig.add_axes((0.1,0.4,0.8,0.5))
ax1.set_title("This is my title")
ax1.set_xlabel(r'\begin{center}X-axis\\*\textit{\small{' + txt + r'}}\end{center}')
ax1.set_ylabel('Y-axis')
ax1.scatter(x,y, c='r')
plt.xlim(0, 1.05)
plt.ylim(0, 2.5)
plt.show()

I did the same thing with the random scatter plot from tacaswell's answer, and here's my result:

One warning: if you tweak this to take input string variables, the strings may not be properly escaped for use with TeX. Escaping LaTeX code is already covered on Stack Overflow, at https://stackoverflow.com/a/25875504/1404311 . I used that directly, and then could take arbitrary xlabels and captions.

Answer 4

as @Nicky V mentioned.

plt.xlabel('''Butterfly contract traded

Note: c1-c2-c3 indicates the position: long 2 times c2 and short c1 anc c3''')

Result:

Answer 5

Another simple solution that I used when I couldn't figure out how to do it in a way that was designed...if that makes sense, was to just add some new lines to the xlabel plt.xlabel("xaxis label\n\n\n\ncaption") This just puts the caption a couple new lines under the x axis which makes it look like a caption even thought it's really part of the xaxis label

Answer 6

I found another trick, using plt.text() you can set the y position below the plot canvas, and it can work as a caption. You can adjust fontsize, colour and even where the line breaks using \n. I am using this method

Answer 7

You can simply use negative position for the y to extend the figure outside the current size.

fig.text(0.5, -0.05, 'some caption', ha='center')

The coordinate is of the figure, where 0,0 means the lower left end, and 1,1 corresponds to the top right.

An example -

import numpy as np
import matplotlib.pyplot as plt

x = np.linspace(0, 1, 100)
y = np.random.rand(100)
plt.scatter(x, y)
fig = plt.gcf()
fig.text(0.5, -0.05, 'A scattered plot with caption', ha='center')