Editors’ Vox is a blog from AGU’s Publications Department.
Researchers are increasingly using computational notebooks to share workflows and data analyses with others. They are also used in the classroom setting, being effective tools to teach data science and other topics. Research computing services often highlight their support of notebooks as a method to interact with their services and facilitate collaboration and sharing (e.g., Princeton, Caltech). As a result, notebooks are fast becoming ubiquitous in research workflows.
Jupyter and RStudio, are two prominent development environments in the computational notebooks space. GitHub, a popular platform for collaborating on and sharing code, including notebooks, currently shows 4 Million code results for “.ipynb”, the file extension used for Jupyter, while the “.rmd” R extension used in, for example, RStudio, has 2.7 Million code results. We have seen a progressive growth in the preservation and citation of software primarily via GitHub and Zenodo, a repository maintained by CERN and OpenAire, climbing from roughly 2 Thousand records in 2014 to 73 Thousand in 2021.
Jupyter Notebooks started out as IPython, a command shell for interactive computing mainly in Python. Since its initial release in 2001, IPython has now developed into the Project Jupyter community whose goal is to “develop open-source software, open standards, and services for interactive computing across dozens of programming languages”. The RStudio integrated development environment for the R statistical programming language was initially released in 2011. Shortly after, R Markdown, an approach to embed R code chunks in Markdown documents, was introduced into the knitr package. Posit (the company previously known as RStudio) recently developed Quarto, the next generation of R Markdown. Quarto is open source with support for code written in R, Python, Julia, and Observable JavaScript.
Notebooks fuse research narratives with data, visualization, and executable code to create an interactive experience.
Notebooks fuse research narratives with data, visualization, and executable code to create an interactive experience that allows others to walk through your computational workflow and analysis. For this reason, notebooks have also been an effective teaching tool in the classroom (e.g. Data8). Advances in notebook development have made it easier to launch a virtual platform, get started with coding, and use them while integrating them into your research workflows (e.g. OpenScapes).
At AGU, we have seen increasing interest in publishing notebooks from groups such as EarthCube, where they feature an annual call to their community to share and publish interactive workflows and analyses of research. Notebooks from the EarthCube calls are published in AGU’s preprint server, Earth and Space Science Open Archive or ESSOAr. In AGU journals, for instance, the Journal of Advances in Modeling Earth Systems (JAMES), editors have highlighted the benefits of sharing notebooks to understand the underlying models presented in papers. The Earth and Space Science journal has also seen interest in publishing notebooks as the research paper itself.
AGU will work with the computational notebooks community to design and implement a publishing workflow for computational notebooks.
AGU recently published two brief guides on publishing notebooks, Erdmann et al. (2021a,b). The guidance was co-developed with community members and continues to evolve. It also inspired the recently funded Alfred P. Sloan Foundation grant to support the advancement of notebooks as a primary research object led by AGU. Guided by a Steering Committee of experts, AGU will work with the computational notebooks community to design and implement a publishing workflow for computational notebooks over the span of 18 months starting in July 2022. Interactive functionality of notebooks is lost during publication, which is primarily still text-based. The project, called Notebooks Now! will look at maintaining the interactivity of notebooks in publishing workflows, effectively documenting reproducible workflows, expanding equitable access to computational research, and extending new forms of contributorship in the wider research community. In addition, this effort includes developing pilot solutions that will elevate notebooks as a format in scholarly publishing. You can learn more about Notebooks Now! at the project website.
In recent years, we have seen the increased use of notebooks in abstracts (presentations and posters) submitted to the AGU’s annual meetings and we are looking at ways to elevate and incentivize their use in our community. Please let us know through this form if you are interested in staying connected with these efforts and/or providing feedback.
—Christopher Erdmann (CErdmann@agu.org, 
0000-0003-2554-180X), American Geophysical Union; Shelley Stall; Brooks Hanson; Laura Lyon; Brian Sedora; Matt Giampoala; Mia Ricci; and AGU’s Notebooks Now! Steering Committee