3d point selection

ipyvolume has a nice implementation of this where there are 3 possible selection modes: rectangle, circle, and lasso (https://github.com/maartenbreddels/ipyvolume/issues/124#issuecomment-376805210)

(GIF via https://towardsdatascience.com/multivolume-rendering-in-jupyter-with-ipyvolume-cross-language-3d-visualization-64389047634a)

This would probably be mostly useful for Dash; manually select a subset of points in a 3d plot and let the Dash app do something with that subset.

[chriddyp]

Another interesting software to look at for inspiration is the 3D scatter plots in https://projector.tensorflow.org/

[seabirdman]

ColorSpace app is a work-in-progress but has some callbacks in a 3D bubble plot that might relate to this discussion. You can access the app here and will see that we show some data about the point on hover and highlight the underlying data upon click. We would like to do more but for now, this seems to be about it and we have moved on to other things. BTW, in the ColorSpace you can toggle from a 3D bubble plot to a color bar chart that has the same callbacks as mentioned above. Write me if the code for these features are of interest.