Outgrowing Your Laptop with Positron
summary
Demonstration shows Positron IDE scaling a Python workflow that reads the Ursa Labs NYC taxi Parquet set (about 36 GB on disk, ~1.5 billion rows) using Ibis + DuckDB: counting rows from metadata is instant, a January-2009 filtered ggplot runs ~20s locally, and loading the entire dataset runs faster on a beefy remote. The built feature Positron remote SSH runs compute and file IO on the remote host, requires no code changes, supports disconnect/reconnect, and is free in Positron desktop; Posit Workbench adds enterprise management.
AI-written summary · about
chapters (8)
- 0:00Introduction
- 1:53Modern tooling strategies
- 6:11Demo: Ibis and lazy evaluation
- 11:52Why Positron excels
- 12:14Remote SSH for compute
- 14:48Demo: remote SSH
- 17:41Posit Workbench for teams
- 19:28Q&A
related talks
21mEnd-to-end data science with the Positron IDE
01h 18mIntroducing Positron, a new data science IDE
45mAnalytics for not so big data with DuckDB
20mData Science in Production: The Way to a Centralized Infrastructure
58mUpdates from Posit
43mThe Column's the limit
33mData preparation and analytics on your laptop with DuckDB
37mDuckDB: Crunching Data Anywhere, From Laptops to Servers