AutoDocish: Automated-ish Dataset Documentation
summary
A practical demo of AutoDocish claims a working, editable pipeline: a command line Python tool that reads CSVs, emits per-file Markdown profiles and a combined JSON, and reports headers, descriptive stats and unique-value counts (examples: 103 unique values found in one column, 35 in sepal length of the Iris CSV). It runs on plain Python 2.7 with no external dependencies, targets CSVs only, and produces editable Markdown templates that are shown to write out correctly.
AI-written summary · about
chapters (12)
- 0:00Introduction
- 1:04Data documentation problems
- 4:19Minimum viable docs
- 6:26Good vs bad docs
- 13:42AutoDocish purpose
- 14:46How AutoDocish works
- 16:33Generated data profiles
- 20:04Technical details
- 21:49Python standard library
- 23:14Future directions
- 24:17Live demo: Iris dataset
- 26:14Q&A
related talks
01h 07mMeet Docling: The “Pandas” for Document AI
50mContinuous and Automated Documentation Docs As Code in Practice
47mDocumentation Is Fine, Don't Worry
46mAsciidoctor: because writing docs does not have to suck
02h 59mDocumenting your code from docstrings to automated builds
18mThanks, I Made It with Quartodoc
28mWriting Docs Is a Chore
01h 25mLearn to Unlock Document Intelligence with Open-Source AI