How are NumPy and SciPy related?

NumPy was originally named scipy.core. NumPy and SciPy are closely related projects. The main reason for the separation is to ensure that the array library (NumPy) is lean and mean, as the bulk of SciPy is not always needed. Also, there was as decision among scientists to retire the array packages numeric (MIT) and numarray (NASA) in favor of scipy.core, and thus it got the name NumPy. SciPy has still not reached 1.0, whereas NumPy is currently released as 1.8.1. NumPy has some facilities for FFT and linear algebra, but not as extensive as SciPy.

Numpy is a foundation on top of which these other ones sit. Itâ€™s a dependency for many things. numpy makes number crunching go faster, and in fewer lines of code. You might use numpy to turn pictures into numbers, apply some algorithm, and convert back again (one use).

scipy is likewise a goldmine of canned types. Use scikit-learn to work through the tutorials that use it, and then to work on your own machine learning (ML) projects with the models it contains.