So, is this the same as this notebook? https://observablehq.com/@lborohfang/rampvis-idiom-integrated-algorithmic-tools-for-visual-ana

Yes it is the same data.

May be say a little more about this to unpick it: "Levels of data granularity are driven by pre-defined relevance criteria." Can you explain that in a little more detail - friendly language!

See comments below. If this is really about moving from a high number of small (pixel) marks that encode data, to a smaller number of larger marks (glyphs or bars) through interactive filtering (selection or zoom/pan, which are essentially spatial filters) then I think that you could say this more clearly. Is that the focus?

It would be great if you could focus on a specific Pixel-based technique ... presumably the one you used in EnsembleVis. Can you cite the paper here? The Keim et al paper could be used in reflection to start thinking about ways to overcome any issues -scalability, etc.

For example, did you use PixelBoosting? did it work?

Do you have permission to reproduce the figures? It would be great if we could explain these to a general scientific audience - what do they mean and show?

I changed the description a little - I hope this is what you mean: "In our visualization each row represents a configuration and each column its parameters." But is this a Scalable Pixel Based visualization? If so - can you show how? I presume this starts with coloured 1 pixel per configurationXparameter, but this isn't shown very clearly. Can you maybe walk us through this change from pixel to glyph based - it's really important and clever! Also - related - we came up with some guidelines for this kind of thing in Sarah Goodwin's thesis. I am sure there is other work, but it may be worth looking at / pointing to this. https://www.gicentre.net/featuredpapers/#/goodwinvisualizing2015/

It feels as though the thing you are testing here is really interactive selection and zoom from pixel to glyph. Maybe that's what "Scalable Pixel Based Visualization" means, but I didn't get this from the outset!

"The goal ..." statement would work very well in the description section.

See comments below and JD's fork with (Potentially) useful edits here. https://observablehq.com/@jsndyks/rampvis-idiom-scalable-pixel-based-visualization You can fork this back and edit further!