multivariate-gaussian.py

# Generate the linear evenly spaced xs and ys
n = 50
three_xs = np.linspace(-5, 5, n)
three_ys = np.linspace(-5, 5, n)

# Generate the values for both the X and Y (two parts to the Z)
three_zs_a = np.array(list(map(lambda x: guass_func(x, 0, 1), three_xs)))
three_zs_b = np.array(list(map(lambda y: guass_func(y, 0, 2), three_ys)))

# Use the Kronecker product of the two https://docs.scipy.org/doc/numpy/reference/generated/numpy.kron.html
three_zs = np.kron(three_zs_a, three_zs_b)

fig = plt.figure(figsize=(15, 10))
ax = fig.add_subplot(111, projection='3d')

# Plot the data by forming into n x n matrices - using tile and repeat we can get the grid specified
ax.plot_surface(np.tile(three_xs, n).reshape((n, n)), 
                np.repeat(three_ys, n).reshape((n, n)), 
                three_zs.reshape((n, n)),
                cmap=cm.coolwarm)
plt.title("Z the joint distribution of X~N(0, 1) and Y~N(0, 2)", y=1.03)
plt.show()