Matplotlib so log axis only has minor tick mark labels at specified points. Also change size of tick labels in colorbar

Question

I am trying to create a plot but I just want the ticklabels to show as shown where the log scale is shown as above. I only want the minor ticklabel for 50, 500 and 2000 to s

Answer 1

1) Use FixedLocator to statically define explicit tick locations.

2) Colorbar cbar will have an ax attribute that will provide access to the usual axis methods including tick formatting.

import numpy as np
import matplotlib.pyplot as plt

fig = plt.figure()
ax = fig.add_subplot(111)
x = np.arange(10,3000,100)
y = np.arange(10,3000,100)
X,Y = np.meshgrid(x,y)
Z = np.random.random(X.shape)*8000000
surf = ax.contourf(X,Y,Z, 8, cmap=plt.cm.jet)
ax.set_ylabel('Log Frequency (Hz)')
ax.set_xlabel('Log Frequency (Hz)')
ax.set_xscale('log')
ax.set_yscale('log')
ax.xaxis.set_minor_formatter(plt.FormatStrFormatter('%d'))
# defining custom minor tick locations:
ax.xaxis.set_minor_locator(plt.FixedLocator([50,500,2000]))
ax.yaxis.set_ticks_position('left')
ax.xaxis.set_ticks_position('bottom')
ax.tick_params(axis='both',reset=False,which='both',length=8,width=2)
cbar = fig.colorbar(surf, shrink=0.5, aspect=20, fraction=.12,pad=.02)
cbar.set_label('Activation',size=18)
# access to cbar tick labels:
cbar.ax.tick_params(labelsize=5) 
plt.show()

Edit

If you want the tick marls, but you want to selectively show the labels, I see nothing wrong with your iteration, except I might use set_visible instead of making the fontsize zero.

You might enjoy finer control using a FuncFormatter where you can use the value or position of the tick to decide whether it gets shown:

def show_only_some(x, pos):
    s = str(int(x))
    if s[0] in ('2','5'):
        return s
    return ''

ax.xaxis.set_minor_formatter(plt.FuncFormatter(show_only_some))

Answer 2

Based on the answer from @Paul I created the following function:

def get_formatter_function(allowed_values, datatype='float'):
    """returns a function, which only allows allowed_values as axis tick labels"""
    def hide_others(value, pos):
        if value in allowed_values:
            if datatype == 'float':
                return value
            elif datatype == 'int':
                return int(value)
        return ''
    return hide_others

Which is a bit more flexible.