When using matplotlib, how do I set the on-screen lengths of the x and y axes to be equal without changing the limits of either axis?

I would like to make the axes of a matplotlib plot a square, and do so by stretching or compressing the scaling of the axes rather than by changing the axis limits. How can this be done?

(In other words, I am seeking a Python analog of the MATLAB command axis square.)

So far, I have tried:

• plt.gca().set_aspect('equal'). This seems to equalize the scaling, rather than the on-screen lengths, of the x and y axes, consistent with the matplotlib documentation. The axes of the resulting plot are not square unless ymax - ymin = xmax - xmin, where xmin, xmax, ymin, and ymax are the axes limits (and this equality does not hold for my plot).

• plt.axis('equal') and plt.axis('scaled'). These seem to provide the results shown and described in this post. In general, the axes of the resulting plot are not square.

If I understand correctly then I think what you're looking for is ax.set_box_aspect(1) (also see here), which will give you square axes with the same limits as the original plot.

Example:

import numpy as np
import matplotlib.pyplot as plt

plt.close("all")

x = np.linspace(0, 50, 200)
y = np.sin(x)

fig, ax = plt.subplots()
ax.plot(x, y)
ax.set_box_aspect(1)
fig.tight_layout()