Lightning Talks
summary
Describes an early but working library that models queries as composable schema-to-schema operations rather than immediate dataframe calls, claiming parity with dplyr style verbs and the ability to emit SQL or run on pandas. Shows a small implementation used internally at winvector LLC, explains operators (extend, join, summarize, window), mentions experiments with Modin and RAPIDS not finished, and admits limited external users but practical internal use.
AI-written summary · about
chapters (8)
- 0:00Lightning Talks format
- 1:15Data Algebra for queries
- 6:58Topic modeling as matrix factorization
- 11:18UX research for dev tools
- 13:06Jupyter in VS Code
- 15:17Scrapbook for notebooks
- 18:31Avoiding pandas copies
- 23:10Q&A
