matplotlib multiple groups with different scale-CodePudding

I have the following script but the resulting plot for the second group is skewed by the first one.

labels = ['CA',  'ME']
Addmitted = [500000, 100]
Applied = [10000000, 1000]

x = np.arange(len(labels))  
width = 0.35  

fig, ax = plt.subplots()
rects1 = ax.bar(x - width/2, Addmitted, width, label='Admitted')
rects2 = ax.bar(x   width/2, Applied, width, label='Applied')

ax.set_xticks(x, labels)
ax.legend()

ax.bar_label(rects1, padding=3)
ax.bar_label(rects2, padding=3)

fig.tight_layout()
plt.ylim(0, 12000000)

plt.show()

CodePudding user response：

A solution could be to use a logarithmic y scale through

However you have to keep in mind that the proportions between rectangles' heights are not linear anymore but logarithmic.

CodePudding user response：

I am posting as an answer but there is probably a better way to do this:

labels = ['CA',  'ME']
Addmitted_l = [500000, 0]
Applied_l = [10000000, 0]

Addmitted_r = [0, 120]
Applied_r = [0, 300]

x = np.arange(len(labels))  
width = 0.2  

fig, ax = plt.subplots()
rects1 = ax.bar(x - width/2, Addmitted_l, width, label='Admitted')
rects2 = ax.bar(x   width/2, Applied_l, width, label='Applied')

ax2 = ax.twinx()
rects3 = ax2.bar(x - width/2, Addmitted_r, width, label='Admitted')
rects4 = ax2.bar(x   width/2, Applied_r, width, label='Applied')

ax.set_xticks(x, labels)
ax2.legend(loc='upper center')

ax.ticklabel_format(style='plain', axis='y')
fig.tight_layout()

plt.show()