In matplotlib, how do you draw R-style axis ticks that point outward from the axes?

Because they are drawn inside the plot area, axis ticks are obscured by the data in many matplotlib plots. A better approach is to draw the ticks extending from the axes outward, as is the default in ggplot, R's plotting system.

In theory, this can be done by redrawing the tick lines with the TICKDOWN and TICKLEFT line-styles for the x-axis and y-axis ticks respectively:

import matplotlib.pyplot as plt
import matplotlib.ticker as mplticker
import matplotlib.lines as 开发者_C百科mpllines

# Create everything, plot some data stored in `x` and `y`
fig = plt.figure()
ax = fig.gca()
plt.plot(x, y)

# Set position and labels of major and minor ticks on the y-axis
# Ignore the details: the point is that there are both major and minor ticks
ax.yaxis.set_major_locator(mplticker.MultipleLocator(1.0))
ax.yaxis.set_minor_locator(mplticker.MultipleLocator(0.5))

ax.xaxis.set_major_locator(mplticker.MultipleLocator(1.0))
ax.xaxis.set_minor_locator(mplticker.MultipleLocator(0.5))

# Try to set the tick markers to extend outward from the axes, R-style
for line in ax.get_xticklines():
    line.set_marker(mpllines.TICKDOWN)

for line in ax.get_yticklines():
    line.set_marker(mpllines.TICKLEFT)

# In real life, we would now move the tick labels farther from the axes so our
# outward-facing ticks don't cover them up

plt.show()

But in practice, that's only half the solution because the get_xticklines and get_yticklines methods return only the major tick lines. The minor ticks remain pointing inward.

What's the work-around for the minor ticks?

In your matplotlib config file, matplotlibrc, you can set:

xtick.direction      : out     # direction: in or out
ytick.direction      : out     # direction: in or out

and this will draw both the major and minor ticks outward by default, like R. For a single program, simply do:

>> from matplotlib import rcParams
>> rcParams['xtick.direction'] = 'out'
>> rcParams['ytick.direction'] = 'out'

You can get the minors in at least two ways:

>>> ax.xaxis.get_ticklines() # the majors
<a list of 20 Line2D ticklines objects>
>>> ax.xaxis.get_ticklines(minor=True) # the minors
<a list of 38 Line2D ticklines objects>
>>> ax.xaxis.get_minorticklines()
<a list of 38 Line2D ticklines objects>

Note that the 38 is because minor tick lines have also been drawn at the "major" locations by the MultipleLocator call.