[FIXED] Axes tick rotation depending on position in Matplotlib

Issue

Calling plt.spines.set_position() appears to be overwriting the tick labels rotation, but not the labels themselves. Before I call .set_position(), rotation= seems to be working as expected:

df = pd.DataFrame({
    'type': ['A', 'B', 'C'],
    'height': [1, 2, 3]
})

fig, ax = plt.subplots(1, figsize=(8,2))

df.plot(kind='bar', ax=ax)

ax.set_xticklabels(['zero', 'one', 'two'], rotation=90)

When then when I call .set_position(), the labels still appear, but they revert to their original rotation:

ax.spines['bottom'].set_position(('outward', 10))

Is there a way to call .set_position() while retaining the label rotation?

Solution

It's simple. Put the set_xticklabels statement after set_position():

df = pd.DataFrame({
    'type': ['A', 'B', 'C'],
    'height': [1, 2, 3]
})

fig, ax = plt.subplots(1, figsize=(8,2))
df.plot(kind='bar', ax=ax)
ax.spines['bottom'].set_position(('outward', 10))
ax.set_xticklabels(['zero', 'one', 'two'], rotation=90)

plt.show()

Answered By - pakpe