3

I'm not finding it possible to add a second supylabel for a right-hand y-axis of a multiplot.

Can anyone please confirm 1) whether or not it can be done and/or 2)provide guidance on how?

I am trying to achieve this: enter image description here

Because there are a variable number of subplots (sometimes an odd number, sometimes even) across the broader project, using subplot-level labelling to label the "middle" subplot would be problematic.

I'm presently accomplishing with figure level text. Which looks fine within python, but the right label gets cut-off by savefig. I can only get it to work if I dummy-in null ax-level y-labels " \n".

nrows = len(dftmp.GroupingCol.unique())
ncols = 1

fig, ax = plt.subplots(nrows=nrows, ncols=ncols, figsize=(14,10), constrained_layout=True,
                sharex=True)

for e, ep in enumerate(dftmp.GroupingCol.unique(), start=1):

    # define a figure axis and plot data
    ax = plt.subplot(nrows, ncols, e)
    dftmp["ValueCol"].loc[dftmp["GroupingCol"]==ep].plot(ax=ax, kind="bar", color=barcolor_lst) #, use_index=False)

    # horizontal reference line (zero change)
    zero_line = plt.axhline(0, color='k', linewidth=0.8)

    # y-axis extent limits
    ax.set_ylim([50*(-1.1), 50*1.1])

    # create right-hand y-axis
    ax2 = ax.twinx()

    # y-axis extent limits
    ax2.set_ylim([200*(-1), 200])

    # null y-label placeholder to accommodate fig-level pseudo-supylabel
    ax2.set_ylabel(" \n") # requires space and newline to work

# create supylabel for left-axis
supy_left = fig.supylabel("Left-hand y-axis super label", fontweight="bold") #, pad = 7)#, fontdict=fontdict) #fontweight='bold')

# use fig-level text as pseudo-supylable for right-axis
fig.text(x=0.97, y=0.5, s="Right-hand y-axis super label\n\n", size=13, fontweight='bold', rotation=270, ha='center', va='center')

# create super-label for x-axis
supx = fig.supxlabel("Bottom super label", fontweight="bold")

In the absence of the fig.text line I tried naming a second supylabel as a different object and the code runs, but doesn't produce the label.

supy_right = fig.supylabel("Cumulative net change (m^3)", fontweight="bold", position=(0.9,0.5))
CreekGeek
  • 1,809
  • 2
  • 14
  • 24

1 Answers1

1

I have found the suplabels to be a little unreliable in many cases, so I resort to low-level tricks in cases like these:

import matplotlib.pyplot as plt

fig = plt.figure(figsize=(4, 4))

# dummy axes 1
ax = fig.add_subplot(1, 1, 1)
ax.set_xticks([])
ax.set_yticks([])
[ax.spines[side].set_visible(False) for side in ('left', 'top', 'right', 'bottom')]
ax.patch.set_visible(False)
ax.set_xlabel('x label', labelpad=30)
ax.set_ylabel('y label left', labelpad=30)

# dummy axes 2 for right ylabel
ax = fig.add_subplot(1, 1, 1)
ax.set_xticks([])
ax.set_yticks([])
[ax.spines[side].set_visible(False) for side in ('left', 'top', 'right', 'bottom')]
ax.patch.set_visible(False)
ax.yaxis.set_label_position('right')
ax.set_ylabel('y label right', labelpad=30)

# actual data axes

num_rows = 4

for i in range(num_rows):
    ax = fig.add_subplot(num_rows, 1, i + 1)
    ...

fig.tight_layout()

enter image description here

You need to adjust the labelpad values according to your liking. The rest can be taken care of by fig.tight_layout() (you might need to specify the rect though).

EDIT: having re-read your question, have you tried increasing the pad_inches value when calling savefig()?

cheersmate
  • 2,385
  • 4
  • 19
  • 32