Scientific Python distributions bundle Python with the most popular libraries. Currently, there are two popular scientific Python distributions providing Python3: Winpython and Anaconda.
Winpython is a unique scientific Python distribution for Windows that allows you to install Python and its accompanying packages on a USB flash drive and move from one Windows machine to another. Installation is absolutely self-contained. The only thing you need to run it from your flash drive is a computer with a reasonably recent version of Windows. Of course, you need a 64 bit Windows to run a 64 bit version of Winpython
Anaconda is a scientific python distribution available for Windows, Mac, and Linux that is provided for free by Continuum Analytics. It is a full-fledged distribution. The only limitation of the free version is that libraries are not optimized with the Intel compiler of the Intel Math Kernel Library (MKL). The commercial version with optimized libraries is available at a reasonable price. The Anaconda distribution offers conda, a python package manager that can be used to download and install additional python packages from Anaconda repositories.
I use Jupyter notebook provided by both distributions for all examples. Installation and starting Jupyter are pretty straightforward, but if you need help you can check out volume 1 of my Data Science in Python series on Amazon.