Sebastian Rejman
Utrecht UniversityExecutable Manuscripts as Jupyter Notebooks: The Next Evolution in Scientific Publishing?
Scientific publishing has evolved significantly over the course of its history: From letters, to society proceedings to digital-only journals. Yet, the manuscript itself has remained largely unchanged over decades. The manuscript itself remains a static pdf. Yet, the digital ecosystem continued to evolve. With faster internet connections, cheaper computing, and easy-to-learn programming languages like Python, the landscape is changing. Platforms like Google Colab allow researchers to run code remotely, and AI tools accelerate coding, making automation accessible to non-expert programmers.
This raises the question: is a static PDF still the best format for scientific publications? Since the 1980s, Mathematica Notebooks have integrated text with executable code, a concept now popularized by Jupyter Notebooks. When combined with open repositories like OSF and GitHub, these notebooks enable fully transparent, reproducible research. Readers can generate the analyses and figures themselves, reuse data, and adapt analysis code—enabling more collaborative studies with larger datasets, which in turn enable the use of machine learning in experimental chemical research.
In this contribution, we provide an introduction to executable manuscripts on the basis of our recent work on the role of external acidity in the catalytic cracking of polyolefins and provide a guide how to get started.
Sebastian Rejman completed his M.Sc. in chemistry from RWTH Aachen University in 2021 focusing on catalysis and computational chemistry. He is currently a PhD candidate at the Inorganic Chemistry and Catalysis Group in Utrecht working on catalytic cracking of polyolefins. Besides his main research, he works on automating experiments and data analysis in heterogeneous catalysis.