Virtual environments#

One Python installation may not be able to simultaneously meet the requirements of many interdependent packages, some of which may require newer and older versions of the same package.

To get around this issue, it is recommended to work within virtual environments. Virtual environments provide an isolated Python installation with a set of required packages for a given project, environments can be activated and deactivated at will.

Virtual environments can be managed with venv or conda.

A complete conda installation including many packages for data science is provided by Anaconda. A lightweight version Miniconda is also available.