Abstract
Researchers are increasingly interested in using Jupyter Notebooks on HPC clusters, science gateways, and compostable platforms. In addition to using notebooks for scientific workflows, developers can use interactive browser controls, called widgets, to allow users to more easily interact with data and without editing code. In addition to simple widgets, such as sliders, checkboxes, and text inputs, more advanced features can be used to build standalone data dashboards and web applications. Many of these widgets and data visualizations are domain independent, though users may discover them in an ad hoc fashion, which can cause an onslaught of tickets to request particular libraries. To avoid frustration faced by users and system administrators alike, we present a recommended environment for scientists and research software engineers developing Jupyter Notebook on cyberinfrastructure platforms.
Nicole Brewer
Site Editor and PhD Student at Arizona State University
Nicole is a PhD student in History and Philosophy of Science at ASU where she is using network analysis and other methods to empirically study the reproducibility of Jupyter Notebooks used in research.