Float vs np.float32

Screen Link: https://app.dataquest.io/m/160/linear-systems/4/representing-an-augmented-matrix-in-numpy

Just a very general question, I used float instead of np.float32 and the answer seemed to work just fine.

What is best practice? Any advantages to using np.float32 instead of float?

Kindly do this: matrix_one.dtype

When dtype=float, you get dtype('float64'), but when dtype=np.float32, you get dtype('float32'). This brings us to how objects are stored in memory: bits and bytes. If you know the maximum and minimum requirements, you can optimize by choosing just the right sub-datatype for storage.

Check this out and this too