Nice detail - feels like 3 user stories, but it's pretty rich and should be a useful for lens for reflection - which worked and how well?

Good - can you like to the actual data? A sample of the table of samples?

I don't want to sound too finicky here, and this is on the right lines, but very broad. It seems to be describing a system rather than a well defined approach? Would it be easier to narrow down to "parallel plots for PSA" or "Hyperslicexplorer"? Would this work? Or can you identify particular concepts or techniques used by Matkovic et al (2008) or Sedlmair et al (2014) to address this problem? E.g. some aspects of Fig 1 (process) or the refinement plots in Fig 8 of Matkovic et al (2008) or one of the strategies (the one you selected - why?) in Section 5 / Table 1 of Sedlmair et al (2014)? This might help you develop a specific angle through which to report the problem and think about your response to it?

thanks! that's really helpful. I think what's best in our case is the linked parallel coordinates views so I'll work with that.

This really helps - as long as linked parallel plots are justified somewhere in the literature.

Are there examples of PSA in other contexts / disciplines that you can comment on? How has the knowledge listed above been used elsewhere?

So, given the change of focus, we need a link or two to linked parallel plots (or at a stretch other attempts at PSA that use parallel plots).

there's also the example above but didn't want to duplicate. if you need more I can probably dig and find more examples

There's a trade-off between clarity and comprehensiveness. I think this is bang on. Great!

It's hard for us to see what you have done without this section!

Which framework? The Sedlmair thing? Be as explicit as you can. Good though that you are referring directly to VIS knowledge!

Where do concepts like "emulators" come from? This is just the kind of thing we want to hear about ... but be clear about sources.

Important - can you be clearer about the framework? This seems to be the key concept that you used to help, but I am not sure which it is! We need to see the parallel plots and have some evidence that they were easy to understand. By whom, for what? Can you comment on this and show examples. And please relate this thinking to the user story - what did they help with in light of the user story?

Very useful. There is something here about when a system becomes useful (later). Initially, HANDS ON non-disruptive interjections seem important. I am thinking around the notion of "Reaction rapid, System slow" - get stuff done quickly messily (reliably) in existing tools to test data, approaches, people ... and then build more accessible sophisticated support around this.

Might narrow user stories help link tasks to techniques?

On Reaction Rapid / System Slow ... Bring task, vis and workflow together over time and inject more specific and sophisticated vis (and integration) as trust is built, successes accumulate, knowledge develops and design pathway emerges.

Really good - interesting and important reflection, but needs work: - what is the explicit thing you are focusing on (framework?): be specific and explicit and keep focussed! - relate reflection to user story - this isn't mentioned much after it is (usefully) introduced - add examples - this is a visual showcase of activity - can you narrow at all? focus? See my comments and see if this stimulates ideas / focus.

Nice improvements - still a couple of things to fix! Some comments added 6th Sept - please check!

I have a runtime error here - do check!

This is nicely written, but the more it can get around to offering a (documented, from the literature) rationale for the linked parallel plots idiom as a solution, the better!

I added a specific paragraph for easier linking below.

This is really nice - very clear! Helpful, informative.

Do delete these instructions!

Nice - coherent, convincing, clear, addresses user story!