pyOpenSci Promoting Open Source Python Software To Support Open Reproducible Science

Python is one of the most commonly used programming languages. However, its general-purpose nature can make it difficult for scientists to use. The sheer size of the Python community means that scientific software is often not always easily discoverable, not consistently documented, and often utilizes inconsistent workflows. Best practices for developing and contributing to software are not commonly taught to scientists. Further, scientists who build useful tools frequently lack the funding and time to publish and maintain their code. This means that good software often is left unshared, undocumented, or unsupported. As a result, many scientists spend valuable time reproducing functionality that already exists, and are less likely to contribute to other efforts. In the age of big data and even bigger scientific challenges, the research community needs open and collaborative workflows for commonly implemented scientific programming tasks that can be built upon. This will in turn increase the rate of scientific discovery. pyOpenSci supports a community around well-documented, discoverable software that can be used to find, download and work with scientific data using standardized, open, and reproducible workflows.

pyOpenSci is modeled after rOpenSci. This model includes community peer-review, and publication of open, easy-to-use, and efficient scientific software. Once a package is accepted it becomes a part of the discoverable pyOpenSci ecosystem. We also build technical capacity by training scientists to contribute to open source software and to build better tools which will increase the long-term sustainability tools. Contributing to software can be intimidating for newcomers, and so we will foster friendly collaboration through our community forum and easy-to-follow guidelines for new contributors. Throughout all of these activities, we promote a culture of accessible software and data workflows that can benefit all researchers using Python.