trevor manz

@manzt
33 Followers
35 Following
20 Posts
phd-ing in boston

videos from SciPy 2024 are live! here's my talk on a vision for composable interactive data visualization with #anywidget

https://youtube.com/watch?v=CjNSP_yQqrc

- YouTube

Enjoy the videos and music you love, upload original content, and share it all with friends, family, and the world on YouTube.

i keep finding fun/interesting workflows that fall out of quak's query-based design

example: define a table subset interactively, copy the SQL, and run it in
@duckdb cli to build unix pipelines

kudos to @Posit for adding support for Jupyter Widgets in Positron IDE!

just fixed a scroll behavior issue, and now quak works swimmingly in VS Code (left) and Positron (right) thanks to #anywidget

quak ๐Ÿฆ† v0.1.3 adds a hover indicator for summary histograms, thanks to @dvdkouril

https://github.com/manzt/quak/releases/tag/v0.1.3

Release v0.1.3 ยท manzt/quak

   ๐Ÿš€ Features Add hovered value label for Histogram plot  -  by @dvdkouril (#32)    ๐Ÿž Bug Fixes Duplicate second row  -  by @manzt in (#36)     View changes on GitHub

GitHub

Thrilled @marimo_io is standardizing on #anywidget for their plugin API! Our projects are really well aligned, pushing widget interactivity beyond Jupyter.

Learn more in our blog post: https://marimo.io/blog/anywidget

Build plugins with anywidget!

Standardizing the plugin API for marimo

couldn't help myself and made a very crude but working #rstats #htmlwidget out of @manzt new quak

https://github.com/timelyportfolio/quak/tree/htmlwidget/htmlwidget

quak/htmlwidget at htmlwidget ยท timelyportfolio/quak

a scalable data profiler for quickly scanning large tables in Jupyter - timelyportfolio/quak

GitHub
@timelyportfolio @dom My vision for anywidget is to standardize frontend ESM for native execution across various backends (languages). Marimo notebooks already support this, bypassing Jupyter widgets. It also works with the deno Jupyter kernel. Excited about anyhtmlwidget as an R adapter. This promises unified frontend packaging, though backend APIs still need maintenance.

The core idea in quak is that all table state is expressed via database queries. User interactions produce SQL, executed lazily at the database level (DuckDB) to refresh data views.

This SQL can also be accessed in Jupyter to materialize data subsets for further analysis.

Excited to introduce quak ๐Ÿฆ† (https://github.com/manzt/quak) a scalable, interactive data profiler built with #anywidget.

- ๐Ÿ–ฑ๏ธ crossfilter & sort millions of rows in real time
- ๐Ÿ”„ supports any Apache Arrow __dataframe__
- โšก powered by Mosaic & DuckDB
- ๐Ÿ““ materialize data subsets back in Jupyter

GitHub - manzt/quak: an anywidget for data that talks like a duck ๐Ÿฆ†

an anywidget for data that talks like a duck ๐Ÿฆ†. Contribute to manzt/quak development by creating an account on GitHub.

GitHub
i made a little website to select and export calendar invites (.ics) for events at the 2024 #Olympic games https://olympicks.xyz
olympicks

export 2024 summer olympic events to your calendar