JupyterLab: The next-generation Jupyter frontend

Description

JupyterLab, the next-generation end-user application for Project Jupyter, will eventually replace the widely used classic Jupyter Notebook. Brian Granger, Chris Colbert, and Ian Rose offer an overview of JupyterLab, which enables users to work with the core building blocks of the classic Jupyter Notebook in a more flexible and integrated manner. Users can arrange multiple notebooks, text editors, terminals, and custom components into multiple tabs and panels, and these components have been carefully designed to enable users to consume them together or separately in order to support novel data-driven workflows. Examples include dragging cells between different notebooks, running a selection of code from a text file in a console/kernel, and real-time rendering of Markdown documents.

JupyterLab is built on top of a robust npm-based extension system that allows anyone to customize JupyterLab and access the public APIs of its core extensions. JupyterLab has been built with a web-first vision, but there is also a native Electron app that provides a simplified user experience. Brian, Chris, and Ian lead a live demo exploring familiar building blocks along with fun new features such as real-time collaboration, interactive rendering of CSV files with millions of rows, GeoJSON support, and integrated Vega/Vega-Lite visualizations.