Matplotlib stretches histogram2d vertically

Question

I'm using this code:

fig = plt.figure(num=2, figsize=(8, 8), dpi=80,
                 facecolor='w', edgecolor='k')
x, y = [xy for xy in zip(*self.pulse_time_distance)]
H, xedges, yedges = np.histogram2d(x, y, bins=(25, 25))
extent = [-50, +50, 0, 10]
plt.imshow(H, extent=extent, interpolation='nearest')
plt.colorbar()

to produce a 2D histogram, however, the plot is stretched vertically and I simply can't figure out how to set its size properly:

Answer 1

Things are "stretched" because you're using imshow. By default, it assumes that you want to display an image where the aspect ratio of the plot will be 1 (in data coordinates).

If you want to disable this behavior, and have the pixels stretch to fill up the plot, just specify aspect="auto".

For example, to reproduce your problem (based on your code snippet):

import matplotlib.pyplot as plt
import numpy as np

# Generate some data
x, y = np.random.random((2, 500))
x *= 10

# Make a 2D histogram
H, xedges, yedges = np.histogram2d(x, y, bins=(25, 25))

# Plot the results
fig, ax = plt.subplots(figsize=(8, 8), dpi=80, facecolor='w', edgecolor='k')

extent = [-50, +50, 0, 10]
im = ax.imshow(H, extent=extent, interpolation='nearest')
fig.colorbar(im)

plt.show()

And we can fix it by just adding aspect="auto" to the imshow call:

import matplotlib.pyplot as plt
import numpy as np

# Generate some data
x, y = np.random.random((2, 500))
x *= 10

# Make a 2D histogram
H, xedges, yedges = np.histogram2d(x, y, bins=(25, 25))

# Plot the results
fig, ax = plt.subplots(figsize=(8, 8), dpi=80, facecolor='w', edgecolor='k')

extent = [-50, +50, 0, 10]
im = ax.imshow(H, extent=extent, interpolation='nearest', aspect='auto')
fig.colorbar(im)

plt.show()

Answer 2

Not sure but I think you should try to modify your historgram2d :

H, xedges, yedges = np.histogram2d(x, y, bins=(25, 25), range=[[-50, 50], [0, 10]])

EDIT : I didn't find how to exactly fix the ratio, but with aspect='auto', matplotlib guess it right:

plt.imshow(hist.T, extent=extent, interpolation='nearest', aspect='auto')