Create An Inset Axes Without Plotting Twice

matplotview.inset_zoom_axes() can be utilized to create inset axes where we don’t have to plot the parent axes data twice.

from matplotlib import cbook
import matplotlib.pyplot as plt
import numpy as np
from matplotview import inset_zoom_axes

def get_demo_image():
    z = cbook.get_sample_data("axes_grid/bivariate_normal.npy", np_load=True)
    # z is a numpy array of 15x15
    return z, (-3, 4, -4, 3)

fig, ax = plt.subplots()

# Make the data...
Z, extent = get_demo_image()
Z2 = np.zeros((150, 150))
ny, nx = Z.shape
Z2[30:30+ny, 30:30+nx] = Z

ax.imshow(Z2, extent=extent, interpolation='nearest', origin="lower")

# Creates an inset axes with automatic view of the parent axes...
axins = inset_zoom_axes(ax, [0.5, 0.5, 0.47, 0.47])
# Set limits to sub region of the original image
x1, x2, y1, y2 = -1.5, -0.9, -2.5, -1.9
axins.set_xlim(x1, x2)
axins.set_ylim(y1, y2)

# Remove the tick labels from the inset axes
axins.set_xticklabels([])
axins.set_yticklabels([])

# Draw the indicator or zoom lines.
ax.indicate_inset_zoom(axins, edgecolor="black")

fig.show()