Jupyter notebooks and hubs for scientific computing

Michael P. Lamoureux, Heather K. Hardeman-Vooys

The journal Nature recently published an article entitled Why Jupyter is data scientists’ computational notebook of choice. With three years of use under our belts, we discuss our experience with Jupyter notebooks, provide guidelines on how to make the transition to these tools for your own research, and present several useful resources to help you make this transition. As an illustrative example, we present a sample notebook recording our research efforts to achieve a 30x speedup in an implementation of standard finite difference code to numerically simulate acoustic waves in a variable velocity field using WebGL on a GPU-powered video card.