Eliciting Best Practices for Collaboration with Computational Notebooks

Despite the widespread adoption of computational notebooks, little is known about best practices for their usage in collaborative contexts. In this paper, we fill gap by eliciting a catalog data science with notebooks. With aim, first look through multivocal literature review. Then, conduct interviews professional scientists to assess awareness these practices. Finally, analysis 1,380 Jupyter notebooks retrieved from Kaggle platform. Findings reveal that experts are mostly aware and tend adopt them daily work. Nonetheless, they do not consistently follow all recommendations as, depending on specific contexts, some deemed unfeasible or counterproductive due lack proper tool support. As such, envision design notebook solutions allow have prioritize exploration rapid prototyping over writing code quality.