Matplotlib - How to remove color bar but keep the heatmap position unchanged

Question

I made a figure with 3 axes in it. Each axis is a heatmap with the same color bar. I want to only keep the colorbar of the 3rd axis and hide the 1st and 2nd colorbar (but keep the heatmap position unchanged). How could I do it?

Here is my code:

fig=plt.figure()
grid = plt.GridSpec(4, 6)

plt.subplot(grid[0:2,0:5])
ax1=sns.heatmap(df_norm, cmap='Reds', vmin=0.05, vmax=0.7)
ax1.axes.get_yaxis().set_visible(False)
ax1.xaxis.tick_top()
ax1.set_xticklabels(col, rotation=90)

plt.subplot(grid[2,0:5])
ax2=sns.heatmap(df_tgfup, cmap='Reds', vmin=0.05, vmax=0.7)
ax2.axes.get_xaxis().set_visible(False)
ax2.axes.get_yaxis().set_visible(False)
ax2.xaxis.tick_top()
ax2.set_xticklabels(col, rotation=90)

plt.subplot(grid[3,0:5])
ax3=sns.heatmap(df_tgfdown, cmap='Reds', vmin=0.05, vmax=0.7)
ax3.axes.get_xaxis().set_visible(False)
ax3.axes.get_yaxis().set_visible(False)
ax3.xaxis.tick_top()
ax3.set_xticklabels(col, rotation=90)

Here is the fig I made:

And here is the fig I want to make:

Answer 1

The colorbars are defined as axes in your figure. You can access them in fig.axes

[<AxesSubplot:>,
 <AxesSubplot:label='<colorbar>'>,
 <AxesSubplot:>,
 <AxesSubplot:label='<colorbar>'>,
 <AxesSubplot:>,
 <AxesSubplot:label='<colorbar>'>]

You can use the set_visible(False) method to hide them:

fig=plt.figure()
grid = plt.GridSpec(4, 6)

df_norm = pd.DataFrame(np.random.randint(0,10,size=(10, 10)), columns=list('ABCDEFGHIJ'))

plt.subplot(grid[0:2,0:5])
ax1=sns.heatmap(df_norm, cmap='Reds', vmin=0.05, vmax=0.7)
ax1.axes.get_yaxis().set_visible(False)
ax1.xaxis.tick_top()
#ax1.set_xticklabels(col, rotation=90)
fig.axes[1].set_visible(False)

plt.subplot(grid[2,0:5])
ax2=sns.heatmap(df_norm, cmap='Reds', vmin=0.05, vmax=0.7)
ax2.axes.get_xaxis().set_visible(False)
ax2.axes.get_yaxis().set_visible(False)
ax2.xaxis.tick_top()
#ax2.set_xticklabels(col, rotation=90)
fig.axes[3].set_visible(False)

plt.subplot(grid[3,0:5])
ax3=sns.heatmap(df_norm, cmap='Reds', vmin=0.05, vmax=0.7)
ax3.axes.get_xaxis().set_visible(False)
ax3.axes.get_yaxis().set_visible(False)
ax3.xaxis.tick_top()
#ax3.set_xticklabels(col, rotation=90)

Answer 2

You can pass in cbar=False argument to sns.heatmap() wherever you wish to hide the color bar of the heatmap.

i.e.,

ax1=sns.heatmap(df_norm, cmap='Reds', cbar=False, vmin=0.05, vmax=0.7)
ax2=sns.heatmap(df_tgfup, cmap='Reds', cbar=False, vmin=0.05, vmax=0.7)

while ax3 will remain unchanged.

Then to keep the position of subplot 1 & 2 unchanged, you can play around with plt.subplots_adjust(right=num); where num is the position of the right edge of the subplots, as a fraction of the figure width which you'll decide for your desirable adjustment of the subplots.