Matplotlib set_yticklabels shifting

Question

Given the following code:

import matplotlib.pyplot as plt
import numpy as np

x = [1.0, 1.1, 2.0, 5.7]
y = np.arange(len(x))
fsize=(2,2)
fig, ax = plt.subplots(1,1,figsize=fsize)
ax.set_yticklabels(['a','b','c','d'])
ax.barh(y,x,align='center',color='grey')
plt.show()

Why are the labels not showing as expected ('a' does not show up and everything is shifted down by 1 place)?

Answer 1

The locator is generating an extra tick on each side (which are not being shown because they is outside the plotted data). Try the following:

>>> ax.get_yticks()
array([-1.,  0.,  1.,  2.,  3.,  4.])

You have a couple of options. You can either hard-code your tick labels to include the extra ticks (which I think is a bad idea):

ax.set_yticklabels(list(' abcd')) # You don't really need 'e'

Or, you can set the ticks to where you want them to be along with the labels:

ax.set_yticks(y)
ax.set_yticklabels(list('abcd'))

A more formal solution to the tick problem would be to set a Locator object on the y-axis. The tick label problem is formally solved by setting the Formatter for the y-axis. That is essentially what is happening under the hood when you call set_yticks and set_yticklabels anyway, but this way you have full control:

from matplotlib.ticker import FixedLocator, FixedFormatter

...

ax.yaxis.set_major_locator(FixedLocator(y))
ax.yaxis.set_major_formatter(FixedFormatter(list('abcd')))