Behaviour change: do not auto-choose V1 as the first variable

[tmelliott]

• Often, the first variable is some kind of an index or ID variable and makes little sense to plot
• Proposal: set Variable 1 to 'none' by default (like the other variables, and in line with how Desktop behaves)

Pros:

• faster landing on Visualize tab (currently it lags while loading that first graph)
• no ugly/unuseful/unnecessary first graph

Cons:

• different behaviour, but not drastically

Thoughts, @chriswild? (Hopefully this sends you an email you can reply to...)

[chriswild]

Hi Tom, As a general rule I like something to appear immediately whenever we have something sensible to show

But I think your argument that the first variable is often/usually not something sensible to show, and it also causes problems (e.g. too many categories)

So I’m convinced.

Best, Chris

From: Tom Elliott @.> Sent: Tuesday, 14 March 2023 1:09 pm To: iNZightVIT/Lite @.> Cc: Chris Wild @.>; Mention @.> Subject: [iNZightVIT/Lite] Behaviour change: do not auto-choose V1 as the first variable (Issue #308)

• Often, the first variable is some kind of an index or ID variable and makes little sense to plot
• Proposal: set Variable 1 to 'none' by default (like the other variables, and in line with how Desktop behaves)

Pros:

• faster landing on Visualize tab (currently it lags while loading that first graph)
• no ugly/unuseful/unnecessary first graph

Cons:

• different behaviour, but not drastically

Thoughts, @chriswildhttps://github.com/chriswild? (Hopefully this sends you an email you can reply to...)

— Reply to this email directly, view it on GitHubhttps://github.com/iNZightVIT/Lite/issues/308, or unsubscribehttps://github.com/notifications/unsubscribe-auth/ADCO3LLWXGJWNRNP6A7YBLLW36ZKLANCNFSM6AAAAAAVZX4N3A. You are receiving this because you were mentioned.Message ID: @.**@.>>

[tmelliott]

Won't change (at this stage) as it will create more issues than it solves.

[tmelliott]

[wrong issue]

[chrk623]

Turns out to be some work. The visualize module is huge and a bit messy. Also, there is a BUG here, currently it uses "none" in variable2 for a placeholder but if there is actually a "none" column in the uploaded dataset, it dosen't work as expected.

[tmelliott]

Yep - there's a lot going on there.. Don't spend more than 2 hours trying to figure this out - if it's more complicated that that, we'll leave it.

As for the bug, would it be easy to relabel the tabs to "- none -" or something less likely to be a variable?

[chrk623]

I think there is a selected = NULL option for selectInput. I'll fix this along with "auto-choose V1" before the new release. If not we can leave it as an enhancement, I can't estimate the time for this.

[tmelliott]

Let's not bother for now - it's too likely it'll have unanticipated side-effects, so we'll work on it after the summer release.