To do sieves in J like the blog post mentioned, the equivalent numpy would be:

>>> arr = np.array([10, 20, 30])

>>> arr[arr > 15]

array([20, 30])