Posted On: Jun 4, 2020

EMR Notebooks is a managed service that provides a full-managed, Jupyter-based notebook to data scientists. A Jupyter kernel provides programming language support in Jupyter. EMR IPython is the default kernel. Additional kernels include R, Julia, and many more. Today we are announcing a new feature that allows data scientists, analysts, and engineers to install and execute custom kernels on the EMR cluster directly from the EMR Notebook. Before this feature, installing custom kernels on a cluster required a multi-step installation process.

Additionally, you can also install Data Science packages directly on the master node of the EMR cluster. Libraries are available to any user running an EMR Notebook attached to the cluster. Before this feature, you could only install notebook-scoped Python libraries. Installing libraries directly on the cluster allows you to take advantage of the compute capacity of the master node of the cluster to execute those libraries.  

To get started with EMR Notebooks, please visit EMR Notebooks page

This feature is available on EMR release version 5.30.0 or later, excluding 6.0.0, in the regions where EMR Notebooks is available.