Conda environment for running Dask applications in Jupyter

To run Dask applications in Jupyter, your Jupyter notebook must use a conda environment with both Dask and Jupyter conda packages.

To create this conda environment, you have two options:
  • Create your own conda environment by uploading a custom yaml file, or by adding the required conda packages to existing conda environments using the IBM® Spectrum Conductor cluster management console.

    T o review a list of the required packages, click Workload > Instance Groups (as you would to create or modify an instance group), click the Dask or Notebooks tab, and click then the Requirements link. This opens a dialog outlining the conda packages required for your Dask or Jupyter version so that you can prepare your conda environment accordingly.

  • Use the provided sample files for Dask and Jupyter, to build a conda environment.

Once you have created the conda environment, you can run Dask applications in Jupyter.