Installing Python Packages on Winstat

SSCC members who use Python on Winstat are welcome to install any packages they need, but need to be mindful that the Winstat environment is not quite the same as a regular Windows PC.

Python distinguishes between packages that are installed globally (meaning all users of a given computer have access to them) and packages that are installed for a particular user. Winstat has all the packages that are part of the Anaconda distribution installed globally, including pandas, matplotlib, numpy, scipy and many others. Individual users cannot install packages globally, or change packages that were installed globally. However, users can install packages for themselves, including updated versions of globally installed packages. These packages will be stored on your U: drive.

Packages should be installed using pip. (Unfortunately, conda is currently not compatible with Winstat.) You can run pip by first starting the Anaconda Prompt, which you can find under Anaconda in the programs list or by typing anaconda in the search box.

This will give you a command line interface. To install a package for yourself type:

pip install --user packagename

where packagename should be replaced by the name of the package you want to install (for example, pip install --user pandas-datareader). The package will be installed in U:\python, which will be created automatically if necessary.

You can install more current versions of globally installed packages with:

pip install --user --upgrade packagename

You should only do so, however, if you need the latest version of a package immediately, as packages can take substantial space in your U: drive. SSCC staff update the Python installation on Winstat each semester.

Last Revised: 9/26/2018