python - getting x,y,z, mean through a 3D data array -

Suppose I have an array that is NxNxN and I want to create an average array that piles every direction. Xy (average on average), xz (average on y), yz (average on x)

I do for XI:

  np.mean (data, Axis = 1, dtype = np.float64)   
 Do I just use the axis = 1 [or 2 or 3] to pile it in each direction?  
  
 
 
 If your NxNxN array is a  numpy.ndarray :  
  xy = NNN means (axis = 2) xz = NNN.mean (axis = 1) yz = NNN.mean (axis = 0)