A Python distribution is a software bundle, which contains a Python interpreter and the Python standard library. Installer programs for common operating systems are a frequent mode of distribution. Many Python distributions also have package managers so that you can install or upgrade various Python packages.
Some of the most popular distributions are listed below. Distributions which are marked as “scientific” are ones which come with IPython, numpy, pandas, and matplotlib, at a minimum. All of the distributions provide at least one integrated development environment (IDE) for free. A Python IDE provides a Python-aware code editor integrated with the ability to run code from that editor.
Scientific: No, but many scientific packages can be added via the package manager
Platform: AIX, HP-UX, Linux, MacOS X, Solaris, Windows
Overview: http://www.activestate.com/activepython
Downloads: http://www.activestate.com/compare-editions
Package List: too many to list - use search page: http://code.activestate.com/pypm/search:/?tab=name
Package Manager: PyPM
IDE: IDLE, Komodo (must be purchased separately from ActiveState)
Note: ActivePython is one of the oldest Python distributions, but is not particularly geared towards science.
Scientific: Yes
Platform: Linux, MacOS X, Windows
Overview: https://store.continuum.io/
Downloads: [Note: Complete distribution is available for free, but requires registration first. Also, academics can get several powerful commercial add-on products for free with proof of affiliation with an educational institution.]
Package List: http://docs.continuum.io/anaconda/pkgs.html
Package Manager: conda
IDE: Spyder
Note: Continuum provides a Python compiler, called Numba, as part of its distribution. This can compile Python code down to machine code and is aware of how to optimize with special consideration for the popular numpy.
Note: Commercial add-on tools are linked against the Intel Math Kernel Library (MKL) for improved numerical performance.
Scientific: Yes
Platform: Linux, MacOS X, Windows
Overview: https://www.enthought.com/products/canopy/
Downloads: [Note: Academics can get the professional version for free by registering for an academic license at https://www.enthought.com/products/canopy/academic/ .]
Package List: https://www.enthought.com/products/canopy/package-index/
Package Manager: Canopy Package Manager
IDE: IDLE, SciTE
Note: Professional version is linked against the Intel Math Kernel Library (MKL) for improved numerical performance.
Scientific: Yes
Platform: Windows
Overview: https://code.google.com/p/pythonxy/
Downloads: https://code.google.com/p/pythonxy/wiki/Downloads
Package List: https://code.google.com/p/pythonxy/wiki/StandardPlugins
IDE: SciTE, Spyder
Other Tools: Console (enhanced Windows command line window), WinMerge (differencing and merging of files on Windows)
Note: A variant of this distribution is also available for Linux; please see pythonxy-linux.
Scientific: Yes
Platform: Windows
Overview: https://code.google.com/p/winpython/
Downloads: https://code.google.com/p/winpython/downloads/list
Package List: https://code.google.com/p/winpython/wiki/PackageIndex
Package Manager: WinPython Package Manager (WPPM)
IDE: SciTE, Spyder
Other Tools: TortoiseHG (Mercurial version control system integrated into Windows Explorer)
Some Python packages may not be a part of some distributions, but contain files which must be compiled (i.e., they are not “pure Python”). As it can be difficult to compile these files, especially on Windows, there exist third-party repositories of precompiled packages.
Overview: http://www.lfd.uci.edu/~gohlke/pythonlibs/
Downloads: http://www.lfd.uci.edu/~gohlke/pythonlibs/
Package List: http://www.lfd.uci.edu/~gohlke/pythonlibs/
Platform: Windows