Implement BoxPlot accounting for a weight column

[caiojp]

Hello, I was using the mark_boxplot but noticed that it is not able to account for a column with weights for the different lines.

Usually, I work with a lot of data, so we aggregate it by some segments and end up having a column with the number of elements inside each segmentation combination.

I would like to implement a way so the statistical values displayed in the chart account for the count.

But I don't really know where to start, I am a computer science student and that would be my first contribution to a project.

From what I understand, the function calls for BoxPlotDef inside core, but from there I can't find the chart logic.

Can someone help me?

[binste]

Altair produces a valid Vega-Lite specification which is then compiled by Vega-Lite to Vega. The actual computation of the values for these box plots happens in Vega (Javascript). See https://altair-viz.github.io/user_guide/internals.html

This feature would need to be implemented in either Vega-Lite or maybe even Vega. I'd suggest that you open a new discussion in the Vega-Lite repo to get inputs from their maintainers if they are interested.

By the way, have you seen https://vegafusion.io/? It can perform calculations, sich as for a box plot, on a server or even a database (duckdb, Snowflake).