Since I was really disappointed with the overly complicated 3D plotting interface of matplotlib, I decided to write a easier interaface ffor my everyday work. It mainly generates the meshgrid data, needed by matplotlib, automatically. So if you have 2D data, that you want to plot as surface, you can simply write

z = generate_my_2d_data()
sane3d.plot_matrix_surface(z)

The same works for scatter, wireframe, contour and contourf.

If you want to scale the axis, you can additionally give the axis tics as Arrays

z = generate_my_2d_data()
x = np.linspace(-10, 20, z.shape[0])
y = np.linspace(-1, 2, z.shape[1])
sane3d.plot_matrix_surface(x, y, z)

All additional parameters are handed to the original plotting function. Thus the following will work as expected.

sane3d.plot_matrix_surface(z, cstride=10, rstride=20)

I also put a notebook, reproducing all of the mplot3d tutorial examples online.