Search code examples
pythonarraysnumpyarray-broadcasting

How do you calculate the means of a 2D numpy array using broadcasting and/or numpy functions?

I have a 2-dimensional numpy array of integers and I would like to compute a 1-dimensional numpy array containing the means of each array of the 2-dimensional array, so for example

array([[1, 2]
       [3, 4]])

would return

array([1.5, 3.5])

Currently, I am using a list comprehension to do this

[sum(i) / len(i) for i in lst]

which works just fine, but I am curious if there is a way to do this using broadcasting and/or numpy functions. That would also be faster for large arrays, which I plan to use this function on. Any insights would be appreciated.

Solution

  • a = np.array([[1, 2], [3, 4]])

    np.mean(a, axis=1)

    This will give you the expected result.

    In a 2-D array, axis=0 indicates the column and axis=1 indicates the row.