Matplotlib: how to remove spacing between a group of subplots

Question

I have a series of pyplot subplots that I've created using a gridspec. They all have an hspace between them, which is fine, except that I would like to keep three of them without any space. Is there a way to do this? Currently, they look like this:

import matplotlib.pyplot as plt
import matplotlib.gridspec as gridspec

fig = plt.figure()
grid_spec = gridspec.GridSpec(nrows=10, ncols=10)
grid_spec.update(hspace=1.5)

ax1 = plt.subplot(grid_spec[0:4, :])
ax2 = plt.subplot(grid_spec[4:7, :], sharex=ax1)

# I would like to group the next 3 together
# so that they are stacked top to bottom and side by side
ax3 = plt.subplot(grid_spec[7:8, :5])
ax4 = plt.subplot(grid_spec[8:, :5], sharex=ax3)
ax5 = plt.subplot(grid_spec[8:, 5:6], sharey=ax4)

plt.show()

I would like them to be arranged like this so I can plot the following 2-D KDE diagram and have the relevant 1-D diagrams above and to the right (roughly displaying this sort of data crudely drawn in paint):

I appreciate any help with this one. Can't seem to find documentation on this sort of thing. Thanks!

Answer 1

You can use mpl_toolkits.axes_grid1.make_axes_locatable to subdivide the area of a subplot of a 3 x 2 grid.

import matplotlib.pyplot as plt
from mpl_toolkits.axes_grid1 import make_axes_locatable

fig = plt.figure()
gs = fig.add_gridspec(nrows=3, ncols=2, hspace=.5,
                      height_ratios=[4, 3, 3], width_ratios=[7, 4])

ax1 = fig.add_subplot(gs[0, :])
ax2 = fig.add_subplot(gs[1, :], sharex=ax1)


ax3 = fig.add_subplot(gs[2, 0])
div = make_axes_locatable(ax3)
ax4 = div.append_axes("top", "40%", pad=0.2, sharex=ax3)
ax5 = div.append_axes("right", "25%", pad=0.2, sharey=ax3)

ax4.tick_params(labelbottom=False)
ax5.tick_params(labelleft=False)

plt.show()

Also, you can create a subgridspec, like

import matplotlib.pyplot as plt
from matplotlib import gridspec

fig = plt.figure()
gs = gridspec.GridSpec(nrows=3, ncols=2, hspace=.5,
                       height_ratios=[4, 3, 3], width_ratios=[7, 4])

ax1 = fig.add_subplot(gs[0, :])
ax2 = fig.add_subplot(gs[1, :], sharex=ax1)

sub_gs = gridspec.GridSpecFromSubplotSpec(2,2, subplot_spec=gs[2,0], hspace=0.3, wspace=0.1,
                                          height_ratios=[1,3], width_ratios=[3,1])
ax3 = fig.add_subplot(sub_gs[1,0])
ax4 = fig.add_subplot(sub_gs[0,0], sharex=ax3)
ax5 = fig.add_subplot(sub_gs[1,1], sharey=ax3)

ax4.tick_params(labelbottom=False)
ax5.tick_params(labelleft=False)

plt.show()

In both cases you will probably want to fine tune the parameters a bit. In general, the matplotlib gridspec tutorial gives a nice overview with many examples on this matter.