Nice transparent densities overlaid

gistfile1.txt

import numpy as np
import matplotlib.pyplot as plt
from matplotlib.colors import LinearSegmentedColormap
from scipy.stats import norm

def plot_transparent_hexbin(ax, x, y, c, color):
    cdict = {'red':   ((0., color[0], color[0]), (1., color[0], color[0])),
             'green': ((0., color[1], color[1]), (1., color[1], color[1])),
             'blue':  ((0., color[2], color[2]), (1., color[2], color[2])),
             'alpha': ((0., 0., 0.), (1., 1., 1.))}

    new_cmap = LinearSegmentedColormap('Custom', cdict)
    plt.register_cmap(cmap=new_cmap)

    ax.hexbin(x.ravel(), y.ravel(), c.ravel(),
              cmap=new_cmap, linewidths=0., edgecolors='none',
              gridsize=50, vmin=0., vmax=np.max(c))

def eval_on_grid(func, xlims=[0., 1.], ylims=[0., 1.], granularity=100):
    X, Y = np.meshgrid(np.linspace(xlims[0], xlims[1], granularity),
                       np.linspace(ylims[0], ylims[1], granularity))
    xy = np.vstack([X.ravel(), Y.ravel()])
    return X, Y, func(xy).reshape(X.shape)

colors = ((1., 0., 0.), (0., 0, 0.), (0., 1., 0.), (1., 1., 0.), (1., 1., 0.),)

plt.figure(1)
ax = plt.subplot()
for i in xrange(5):
    density_func = lambda x: norm.pdf(x[0, :] - i * 0.1, x[1, :], 1.0 / (0.9 * ( i + 10)))
    X, Y, C = eval_on_grid(density_func)
    plot_transparent_hexbin(ax, X, Y, C, colors[i])

plt.show()

cool code