Making better scatter plot: sort data points for scatter plots to plot first most frequent data

When plotting scatter plot, less frequent data are hidden by the most frequent ones, leading to very crowded scatter plot, reducing the readability of the plot:

%pylab inline
data = numpy.genfromtxt('test.txt')
x = data[:,0]
y = data[:,1]
z = data[:,2]
scatter(x,y,c=z)
colorbar()

The function below sort scatter data from the less frequent one to the most frequent using a 3D histogram:

data_sorted = sort_scatter_data(data)
x = data_sorted[:,0]
y = data_sorted[:,1]
z = data_sorted[:,2]
scatter(x,y,c=z)
colorbar()

The resulting graph shows a larger range of z-values depicted by the colorbar: