Interactive Crossfilter across different plot types and data aggregations

Hello,

first of all, thank you for creating bokeh. It’s a really great library and helpful in a lot of use cases.

Recently, I wanted to create an interactive crossfilter with bokeh.

With interactive crossfilter across different plot types I mean something like this:

  1. https://github.com/uwdata/falcon

  2. https://altair-viz.github.io/gallery/interactive_layered_crossfilter.html

So, the selection is based on rows of the underlying data but the linked plots show different views/aggregations of the underlying data.

When I was looking at the bokeh documentation I found the following example for Linked Brushing:

  1. https://bokeh.pydata.org/en/latest/docs/user_guide/interaction/linking.html#userguide-interaction-linking

However, the ColumnDataSource refers to the same dataframe and only refers to another column. It does not show a different aggregation of the dataframe.

Is it possible to create something like 1) or 2) with bokeh? And if so: how is it possible?

Any hints are highly appreciated and I am glad to provide a new example for the gallery if we can get this to work.

Best regards,

Florian