Enhancing computational notebooks with code+data space versioning

There is a significant gap between how people explore data and how Jupyter-like computational notebooks are designed. People explore data nonlinearly, using execution undos, branching, and/or complete reverts, whereas computational notebooks are designed for sequential exploration only. Recent works like ForkIt are still insufficient to support these multiple modes of nonlinear exploration in a unified way. In this work, we address the challenge by proposing two-dimensional code+data space versioning for computational notebooks and verifying its effectiveness using our prototype, Kishuboard, which seamlessly integrates with Jupyter. By adjusting code and data knobs, users of Kishuboard can intuitively manage the state of computational notebooks in a flexible way, thereby achieving both execution rollbacks and checkouts across complex multi-branch exploration history. Moreover, this two-dimensional versioning mechanism can easily be presented along with a friendly one-dimensional history. Human-subject and LLM-agent-based studies indicate that Kishuboard can significantly enhance user productivity in various data science tasks

Name thesis:degree_name

M.S.

Level thesis:degree_level

Thesis

Discipline thesis:degree_discipline

Computer Science

Grantor

University of Illinois Urbana-Champaign

Year dc:date

2025

Author dc:creator

• Fang, Hanxi

Contributors dc:contributor

• Park, Yongjoo

• Computational Notebooks
• Version Control Systems
• Code+data Version Control
• Notebook Kernel State Checkpoints
• Interactive Data Science Checkpoints
• Version Control User Interfaces
• Ai-agent

Statement dc:rights

• Copyright 2025 Hanxi Fang

Language dc:language

en, eng

Handle dc:identifier

https://hdl.handle.net/2142/129642

OAI record oai:identifier

https://www.ideals.illinois.edu/oai-pmh?verb=GetRecord&metadataPrefix=etdms&identifier=oai%3Awww.ideals.illinois.edu%3A2142%2F129642