Reproducible, reliable, reusable analyses with BinderHub on Cloud

Sarah Gibson, Alan Turing Institute

We aim to support the research across all fields using cloud computing to provide reproducible, reliable, reusable analyses. We plan to achieve this through capturing computing environments without installation, facilitating access to advanced resources and handling private repositories.

Binder/BinderHub is a cloud platform that captures the computing environment of a Git repository within a Docker image. It provides an interactive browser where code can be executed in the developer’s environment without requiring installation. BinderHubs can be built and hosted anywhere in minimal installation time. Building a locally hosted BinderHub is advantageous for using private repositories and accessing a greater variety of computing resources.

By harnessing cloud computing and a BinderHub with expanded capabilities, scientific analyses can become more accessible and transparent to the wider community. We can improve reproducibility by making it easy to showcase and re-run analyses, improve reliability by making it easy to review analyses and catch bugs early in the development process without installation, and improve the reusability of the code by capturing the computing environment.