Mosaic v0.12 is out: database-powered scalable, interactive visualization! 📈 One new addition is support for dynamic changes in the backing data. Move between smaller and larger samples to balance speed and comprehensive coverage.
Comments
Log in with your Bluesky account to leave a comment
Mosaic optimizes queries to a backing DuckDB instance - either in the browser, on a server, or in a Jupyter notebook kernel - to support fast interactive updates. For more, see https://idl.uw.edu/mosaic/
I'm wondering what the relationship between Observable and mosaic looks like.
The mosaic site is obviously built with Observable Framework, and the visualizations look like generated by Plot, but mosaic also seems to support Vega-Lite and other libraries.
Mosaic is a framework for linking interactive views and backing databases. Examples include input widgets and vgplot, an interactive grammar of graphics built on Observable Plot, but you can also build and integrate custom components. The Mosaic core handles database querying and linked selections.
Also, FWIW the Mosaic site is built using VitePress, not Observable Framework. For more explanation of Mosaic, start here: http://idl.uw.edu/mosaic/what-is-mosaic/
My understanding so far is that mosaic adds the possibility to effectively cross-filter the data, which makes it possible that multiple charts can filter the data at the same time.
Mosaic serves as an architectural layer that provides (1) access to a backing database and a variety of query optimizations, and (2) a shared selection abstraction that enables integrated filtering across data views (visualizations, tables, input widgets, etc.).
Wonderful! I have a tiny stats question: Is this showing a regression line of points just within that window? If so, could that "local view" make overall trends look less dramatic than they really are? This interactive visualization shows the effect I'm wondering about: https://claude.site/artifacts/d221586e-fe1e-4954-ac2a-43cb6164eff8
Yes, the local regression is limited to points falling within the brush. Depending on the data and analysis task, there may be utility in seeing localized trends (a simple parametric distribution is a poor model in many cases) but of course statistical inference pitfalls abound. Use with caution 😊
(And in case it’s interesting to you: underneath the hood Mosaic automatically pre-aggregates sufficient statistics, binned at the pixel level, to compute updated regressions on the fly without having to touch the original, large dataset.)
Comments
The mosaic site is obviously built with Observable Framework, and the visualizations look like generated by Plot, but mosaic also seems to support Vega-Lite and other libraries.
Is that about correct?
https://claude.site/artifacts/d221586e-fe1e-4954-ac2a-43cb6164eff8