Composing Flexibly-Organized Step-by-Step Tutorials from Linked Source Code, Snippets, and Outputs

Head, Andrew; Jiang, Jason Zheng; Smith, James; Hearst, Marti A.; Hartmann, Björn

doi:10.1145/3313831.3376798

Cited by 25 publications

(13 citation statements)

References 5 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…G1: Provide suitable EDA examples in context. Because real-world data problems are often vague and ill-defined, data scientists usually search for existing EDA notebooks online to learn how others address similar issues [20,33,49]. Our E1 and E2 confirmed this as well (R3).…”

Section: Design Goalssupporting

confidence: 65%

“…Being able to provide some suggestions about subsequent exploration steps can be significantly helpful, not only in data manipulation [2,59] but also in visualization generation [36,56,57]. This is essential for novices to learn and get familiar with a large number of methods/APIs available in data science toolkits (e.g., pandas) [18,20], thus helping them form rationale for the next steps (R3). While completely automating the whole EDA process without any limitation is not possible, the system should offer some level of suggestion, also in-situ, such as the operations that are likely to use next based on the current code sequence (R2).…”

Section: Design Goalsmentioning

confidence: 99%

“…Indeed, people learn programming from the widely-available examples and tutorials online, and code search is one frequent and critical activity for developers [7,43,47,61]. This is the same for EDA and using notebooks [20,49]. When data scientists, especially novices, start to investigate their data or are stuck at some problems, they usually look for EDA notebooks online to learn how others approach the same or similar problem, check the APIs, models, and metrics that others have used, and get inspirations for performing EDA themselves [44,49].…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

EDAssistant: Supporting Exploratory Data Analysis in Computational Notebooks with In-Situ Code Search and Recommendation

Li¹,

Zhang²,

Leung³

et al. 2021

Preprint

View full text Add to dashboard Cite

Using computational notebooks (e.g., Jupyter Notebook), data scientists rationalize their exploratory data analysis (EDA) based on their prior experience and external knowledge such as online examples. For novices or data scientists who lack specific knowledge about the dataset or problem to investigate, effectively obtaining and understanding the external information is critical to carry out EDA. This paper presents EDAssistant, a JupyterLab extension that supports EDA with in-situ search of example notebooks and recommendation of useful APIs, powered by novel interactive visualization of search results. The code search and recommendation are enabled by state-of-the-art machine learning models, trained on a large corpus of EDA notebooks collected online. A user study is conducted to investigate both EDAssistant and data scientists' current practice (i.e., using external search engines). The results demonstrate the effectiveness and usefulness of EDAssistant, and participants appreciated its smooth and in-context support of EDA. We also report several design implications regarding code recommendation tools.

show abstract

Section: Design Goalssupporting

confidence: 65%

Section: Design Goalsmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

EDAssistant: Supporting Exploratory Data Analysis in Computational Notebooks with In-Situ Code Search and Recommendation

Li¹,

Zhang²,

Leung³

et al. 2021

Preprint

View full text Add to dashboard Cite

show abstract

“…3.3.4 Presentation (10). We also found ten code blocks that presented ancillary information, such as showing a URL.…”

Section: Instruction Output (93)mentioning

confidence: 87%

“…Tutorial takers stumble when tutorials contain missing dependencies or steps [23], do not explain unexpected errors or side effects, and have unclear adaption paths for tailoring content to different goals [18]. Head et al [10] interviewed 12 tutorial creators and discovered pain points related to duplicate instructions and composing and reusing code fragments from a working example. Furthermore, the authors created an interactive tutorial author tool, Torii, which allowed authors to split, annotate, and link tutorial content.…”

Section: Related Workmentioning

confidence: 99%