NoteToon: exploring the effects of visual narratives within the notebook interface to improve communication

Abstract

Computational notebooks help data scientists analyze and visualize datasets, and share analysis procedures and outputs. However, notebooks typically join code (e.g., Python scripts), notes, and outputs (e.g., tables, graphs). The combination of disparate materials is known to hinder notebooks' comprehension, making it difficult for scientists to collaborate with other scientists unfamiliar with the dataset. To mitigate this problem, we introduce NoteToon, a JupyterLab extension that enables conversion of notebooks into 'data comics.' NoteToon provides a simplified view of a Jupyter notebook, highlighting the most significant results while promoting interactive and free exploration of the dataset. This paper presents a formative study that motivated the system, its implementation, and an evaluation with 12 users, showing the effectiveness of the produced comics. We discuss how our findings inform the future design of interfaces for computational notebooks and features to assist various collaborators.