Source: Deep Learning on Medium
Python Numpy handbook for every data scientist
I think one of the most useful scientific libraries is Numpy. Perhaps, over 90% of data scientists and Python developers use Numpy.
Originally Python is not for numeric computing but many scientific and engineering communities are interested in it.
In 1995, Matrix-SIG was founded to define an array computing package. Firstly, Guido van Rossum who is the creator of Python programming language and was a developer at Google and Dropbox implemented indexing syntax to make an array easy to use.
In 1997, Jim Fulton and Jim Hugunin made “Numeric” which is the implementation of a matrix package. And also, another replacement which is called “Numarray” to “Numeric” was created. Even though there was Numarray as a replacement to the Numeric both packages were used. Because Numarray had faster operations for large arrays but was slower than Numeric on small ones.
In 2005, Travis Oliphant unified two packages of Numarray and Numeric as well as related communities. The first NumPy was separated from SciPy and the official NumPy 1.0 released in 2006.
Nowadays, Numpy is essential to various scientific research areas as well as Python applications. This is very useful but it’s very confusing to understand how it works for slicing, indexing, and array manipulating.
I hope these simple sample codes will be helpful to readers.