Loading...
Search for:
kaggle
0.049 seconds
A Solution Towards Automatically Assessing Jupyter Notebooks Code Comprehension on Kaggle
, M.Sc. Thesis Sharif University of Technology ; Heydarnoori, Abbas (Supervisor)
Abstract
A Jupyter Notebook is composed of richly formatted cells containing both markdown and code, making it a primary coding environment for data scientists. While the advantage of seamlessly integrating code, text, and output in these notebooks is evident, criticisms have arisen due to unexpected output and the promotion of poor coding patterns, raising concerns about maintainability and reusability of these notebooks. The unique characteristics of Jupyter Notebooks, combining code and text, pose challenges for traditional methods that rely solely on source code analysis to evaluate comprehension. Notebook-related metrics such as number of markdown cells and developer-related such as performance...