Nice that you have direct links, but "404 Page Not found" Can you fix this?

".. parallel axes representing variables and their ranges"? Can you also cite Inselberg's paper on this - it's the standard source. The EuroVIS STAR should be a useful source too. It's quite a good starting point. Heinrich, J. and Weiskopf, D., 2013, May. State of the Art of Parallel Coordinates. In Eurographics (State of the Art Reports) (pp. 95-116).

Sorry - above comment should relate to below section!

This is a bit confusing. The final sentence is fine, but the rest is a description of the method - so add it to section 3. Can you consider adding Inselberg's paper on PCPs here too - it's the standard source. The EuroVIS STAR should be a useful source too. It's quite a good starting point. Heinrich, J. and Weiskopf, D., 2013, May. State of the Art of Parallel Coordinates. In Eurographics (State of the Art Reports) (pp. 95-116).

"Parallel Coordinates Plots"?

That seems to be the focus, but see discussion below. You may want to narrow the focus.

Remember - we are trying to get scientists on side here. They won't know what "the cars data set" is. Please explain fully what this shows and how.

I don't understand why the smart brushing is important or relevant here, but if you used it then this could be a useful idiom around which to do some thinking! As an epidemiologist I want to select an important subset of 160,000 multivariate model parameter configurations So that I can understand error, uncertainty and parameter relationships to inform modelling Does that make sense? Did it work in this case?

OK - much better - can you add an interpretation? "The plot allows us to see the expected strong inverse relationships between weight and acceleration, and economy (MPG) and number of cylinders (CYL). Bit to find out more we need to highlight cars of particular yea or origin to see how they vary across multiple variable axes". "Parallel coordinates plots show 'signatures' across multiple variables but are highly dependent upon axis ordering. It is, for example difficult to relate number of cylinders (CYL) with acceleration here, but would be straightforward if the axes were adjacent. Once again, this highlights the need for interaction."

Or use and introduce the term "brushing" which is key to your **excellent** example that follows!

Good description of what you have done. But what dies it enable you to know in terms of the user story? Can you add something?

This is great - very clear. Is there any way you can relate it to the user story? Finding a maximum is pretty easy (it's a database query) and univariate. Can you interpret this graphic in ways that take advantage of the parallel plot and address the user story on error, uncertainty and comparison?

There is certainly room for some work here in thinking about how well the idiom addressed the user story and what was learned in trying to use it for this purpose. I hope the group can get together to do some further thinking!

For example, does the smart brushing work with this many axes and 25M data points? If not, what did you have to do and what did this cost?

Delete this

Can you add some recommendations that come from your experience of using PCPs in RAMPVIS?