To Notebook or Not to Notebook: Multilingual Workflows for Data Analysis
summary
Notebooks enable combining Python, R and SQL so a workflow can run SQL, Python, and R in one notebook using kernels and extensions like R2Py or Reticulate and renderable documents like Quarto, but multilingual notebooks introduce harder debugging, dependency bloat, performance overhead, and increased team knowledge needs. Historical scale: Jupyter hits ~2.5M notebooks on GitHub by 2018 and nearly 10M later; a provided repo is a runnable boilerplate and Positron ships an IDE with data explorer, variables pane, and Positron assistant edits notebooks live.
AI-written summary · about
chapters (13)
- 0:00Introduction
- 1:15Talk title and context
- 5:07Jupyter Notebooks anatomy
- 9:23Jupyter Notebooks history
- 10:44Notebook controversies
- 11:31Why data scientists love notebooks
- 13:31Data scientists as multilingual
- 15:31Multilingual notebook workflows
- 17:32Live multilingual notebook demo
- 19:25Multilingual workflow challenges
- 21:00Notebook editors for multilingual
- 23:06AI in notebook tools
- 23:55Closing thoughts
related talks
43mJupyter: Notebooks in Multiple Languages for Data Science
49mBest tools for Jupyter notebooks
34mReactive Notebooks for Python
45mData Science Apps: Beyond Notebooks
29mOh no! I think my project is outgrowing my Jupyter notebook. How do I survive?
39mClean Code in Jupyter notebooks, using Python
34mNotebooker: Production and Scheduling for your Jupyter Notebooks
24mJupyter Notebooks for Ops