About

Here is a description of my preferred way of working with Jupyter (being either Notebook or Lab). The plan is to basically keep the Jupyter installation within the base environment. Each additional virtual environment is then exposed via an IPython kernel and used via Jupyter.

Procedure

  1. install anaconda.

    install in user mode and not in admin mode.

  2. start the Anaconda Prompt (anaconda3) using start menu.

    this should automatically activate the base environment.

  3. create a conda environment named (for example) data-science-env.

    conda create --name data-science-env python=3.8

  4. activate the conda environment.

    conda activate data-science-env

  5. install any required package using conda or pip.

    pip install scikit-learn or conda install scikit-learn

  6. create a notebook kernel connected with the conda environment just created.

    python -m ipykernel install --user --name data-science-env --display-name data-science-env-kernel

  1. use the base environment to start jupyter-notebook or jupyter-lab.

    You should find a kernel named data-science-env-kernel.