If you are anything like me you start multiple data science projects. To make things a little easier I make multiple conda environments. Conda does a decent job at reducing the space used my the packages[^symlinks]. But over time the conda folders can become boated. It’s a good idea to do this regularly.
conda clean --yes --all
The all flag clears packages, and tarball used in the installation process.
You can also add a dry-run flag:
conda clean --all --dry-run
to check the size of unused packages. Can be quite large, if you have deleted/updated any conda environment!