Integrate with tabular functionality

[jo-mueller]

Hi @dstansby ,

I was recently discussing with @zoccoler about how to further improving napari-matplotlib. One thing that appeared to be a natural extension for plotting functionality would be the support for some sort of tabular interaction: Clicking on column heads could trigger the display of feature scatter/histogram plots, similarly for metadata. We could introduce standardized column headers (label already exists) such as frame for the time dimension, etc.

Now such things obviously already exist (e.g., in the scope of napari-spreadsheet) and it would be superfluous effort to create this functionality from scratch. Another place where tabular functionality exists (and is actively used) is napari-skimage-regionprops (see code here).

Pitch:

In order to make the most of both functionality (current napari-matplotlib and table from napari-skimage-regionprops), we suggest to join efforts and merge these two pieces of code under one hood. This would greatly enhance the reach of napari-matplotlib, its usability and spread the effort of maintenance over more shoulders.

Would love to hear your thoughts!

Also tagging @haesleinhuepf as the maintainer of napari-skimage-regionprops

[dstansby]

I'm not quite sure from the description above what you're trying to achieve with tabular functionality. Could you provide some example workflows that involve 1) the napari viewer 2) a table of data and 3) a 2D plot (which is where napari-matplotlib comes in), and hopefully from those we can explore what napari-matplotlib could provide to support those workflows.