matplotlib: log transform counts in hist2d

Question

Is there a simple way to get log transformed counts when plotting a two dimensional histogram in matplotlib? Unlike the pyplot.hist method, the pyplot.hist2d method does not seem to have a log parameter.

Currently I'm doing the following:

import numpy as np
import matplotlib as mpl
import matplotlib.pylab as plt

matrix, *opt = np.histogram2d(x, y)
img = plt.imshow(matrix, norm = mpl.colors.LogNorm(), cmap = mpl.cm.gray, 
                 interpolation="None")

Which plots the expected histogram, but the axis labels show the indices of the bins and thus not the expected value.

Answer 1

It's kind of embarrassing, but the answer to my question is actually in the docstring of the corresponding code:

Notes
-----
    Rendering the histogram with a logarithmic color scale is
    accomplished by passing a :class:`colors.LogNorm` instance to
    the *norm* keyword argument. Likewise, power-law normalization
    (similar in effect to gamma correction) can be accomplished with
    :class:`colors.PowerNorm`.

So this works:

import matplotlib as mpl
import matplotlib.pylab as plt
par = plt.hist2d(x, y, norm=mpl.colors.LogNorm(), cmap=mpl.cm.gray)