How do I make a matplotlib scatter plot square?

Question

In gnuplot I can do this to get a square plot:

set size square

What is the equivalent in matplotlib? I have tried this:

import matplotlib
matplotlib.use('Agg')
import matplotlib.pyplot as plt
plt.rcParams['backend'] = 'TkAgg'
x = [0, 0.2, 0.4, 0.6, 0.8]
y = [0, 0.5, 1, 1.5, 2.0]
colors = ['k']*len(x)
plt.scatter(x, y, c=colors, alpha=0.5)
plt.axes().set_aspect('equal', adjustable='datalim')
plt.xlim((0,2))
plt.ylim((0,2))
plt.grid(b=True, which='major', color='k', linestyle='--')
plt.savefig('{}.png'.format(rsID), dpi=600)
plt.close()
plt.clf()

I get a square grid, but the plot itself is not square. How do I make the x range go from 0 to 2 and make the plot square?

Rybinsk

Answer 1

You can do it like this:

import matplotlib.pyplot as plt

fig, ax = plt.subplots()
x = [0, 0.2, 0.4, 0.6, 0.8]
y = [0, 0.5, 1, 1.5, 2.0]
colors = ['k']*len(x)
ax.scatter(x, y, c=colors, alpha=0.5)
ax.set_xlim((0,2))
ax.set_ylim((0,2))
x0,x1 = ax.get_xlim()
y0,y1 = ax.get_ylim()
ax.set_aspect(abs(x1-x0)/abs(y1-y0))
ax.grid(b=True, which='major', color='k', linestyle='--')
fig.savefig('test.png', dpi=600)
plt.close(fig)

Superhighway

At least by now (February 2020), you may use set_aspect('equal') to obtain the «same scaling from data to plot units for x and y» to offer the same result. – Uriah

Answer 2

Set the size in the figure:

import matplotlib
matplotlib.use('Agg')
import matplotlib.pyplot as plt
plt.rcParams['backend'] = 'TkAgg'
x = [0, 0.2, 0.4, 0.6, 0.8]
y = [0, 0.5, 1, 1.5, 2.0]
colors = ['k']*len(x)
fig = plt.figure(figsize=(6,6)) # default is (8,6)
ax = fig.add_subplot(111, aspect='equal')
ax.scatter(x, y, c=colors, alpha=0.5)
ax.set_xlim((0,2))
ax.set_ylim((0,2))
ax.grid(b=True, which='major', color='k', linestyle='--')

Twelfth

Would the individual who down-voted the answer care to explain why? – Twelfth

It wasn't me, but I think what you did only sets the canvas size to equal, it doesn't set the actual plot area to being square. It can be messed up by the axis and tick labels and so on, because matplotlib will shrink the plot area to fit those things onto the canvas. – Kraemer