Custom positioning of x-ticks & labels

Question

For a simple bar chart, I want some of my x-tick labels to be below the x-axis (for positive y-values) while others are above the x-axis (for negative y-values).

I see, per the docs, that a string can be passed as an argument to matplotlib.axes.Axes.set_xticklabel to change vertical alignment, but this is a blanket application.

How would I achieve arbitrary positioning of particular x-tick labels?

Answer 1

You can modify the position / rotation of the labels in a for loop. Initially, I thought changing the rotation point would be enough, but the labels are anchored at the tip of the ticks, so the result was sub-optimal.

This is my code. You may want to tweak the y position in plus or minus to get the alignment that you want:

N = 10
h = np.linspace(-10,10,N)
xticks = [f'test #{n}' for n in range(N)]

fig, ax = plt.subplots()
ax.spines['bottom'].set_position('zero')
ax.bar(range(N),h)
ax.set_xticks(range(N))
ax.set_xticklabels(xticks)

for i,(g,t) in enumerate(zip(h,ax.get_xticklabels())):
    if g<0:
        t.set_ha('left')
        t.set_va('bottom')
    else:
        t.set_ha('right')
        t.set_va('top')
    t.set_rotation_mode('anchor')
    t.set_rotation(45)
    t.set_transform(ax.transData)
    t.set_position((i,0))