How Jupyter makes experimental and computational collaborations easy

Scientific research thrives on collaborations between computational and experimental groups, who work together to solve problems using their separate expertise. Zach Sailer highlights how tools like the Jupyter Notebook, JupyterHub, and ipywidgets can be used to make these collaborations smoother and more effective.

Collaboration between computational and experimental research groups is foundational to science. The most effective collaboration involves experimentalists gathering data, computational scientists developing code to analyze that data, and the two groups working together to interpret the results. However, computational scientists recognize this can be challenging, as their experimental collaborators do not find staring at code quite as helpful. If you want a headache-free collaboration between experimental and computational research groups, you need to get code out of the way. How do you develop simple-to-repeat, accessible programs that make your collaborators happy? Zach Sailer highlights various ways in which the Jupyter ecosystem has improved collaboration with experimental groups and shares a real-world example in which computational scientists worked with experimentalists in Australia to tackle the problem of predicting drug resistance in unknown malarial strains. Zach explains how the team used JupyterHub to host a private, centralized server that was shared with their collaborators, making uploading and downloading data and analyses simple and secure, how they used the Jupyter Notebook and ipywidgets to make their computational analyses more accessible, interactive, and reproducible for their collaborators, and how they openly shared Jupyter notebooks upon publication, using external web services like GitHub and Binder.