Plotly express with multiple dataframe variables and color mapping looses variable mapping

[rjvtww]

I am using plotly express by passing y variables as a list of dataframe names, which works perfectly fine and gives the following output:

px.scatter(df, x=df.index, y=yvars, title='Scatter Matrix')

The problem is if I add a color map variable, all the y variables are passed together, and I can no longer tell/select the plotted variables:

px.scatter(df, x=df.index, y=yvars, title='Scatter Matrix', color=categorychoice)

I made sure it was the case by forcing the legend view, as seen in the screenshot.

This is problematic of course as I lose the variables information and its all just one single var.

Thanks in advance!

Best,

Rodolfo

[nicolaskruchten]

Can you describe the output you'd like to see in this case?

[rjvtww]

I would like to retain the information of each plotted Y variable. In the above case, being able to hide/unhide each of plotted vars, as I am able in the first example without color mapping.

[nicolaskruchten]

OK, I can kind of see that, but you want them to be colored based on the same continuous scale?

[rjvtww]

Precisely.

…

On Mon, Nov 16, 2020, 7:46 PM Nicolas Kruchten ***@***.***> wrote: OK, I can kind of see that, but you want them to be colored based on the same continuous scale? — You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub <#2901 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AQVHMDKOGV3IA5CMVFR6JSTSQFXPJANCNFSM4TUL65UQ> .

[nicolaskruchten]

OK, so in px.line() we have the concept of a line_group which does roughly this: it splits the traces, mostly because otherwise you get end-to-start lines where you wouldn't want them. It seems like you'd want to have an equivalent scatter_group argument here... Maybe similar to the bar_group concept here #2657

[rjvtww]

I understand that the line_group concept its to avoid the end-to-start lines, which is very reasonable. But in my case, I feel that the native behaviour should be to maintain the independence of each variable. There is no reason why only by passing a color mapping variable, the behaviour changes.

…

On Mon, Nov 16, 2020 at 7:58 PM Nicolas Kruchten ***@***.***> wrote: OK, so in px.line() we have the concept of a line_group which does roughly this: it splits the traces, mostly because otherwise you get end-to-start lines where you wouldn't want them. It seems like you'd want to have an equivalent scatter_group argument here... Maybe similar to the bar_group concept here #2657 <#2657> — You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub <#2901 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AQVHMDPZD73NBG4T6XIXIR3SQFY67ANCNFSM4TUL65UQ> .

-- R.

[nicolaskruchten]

Well, there is a reason, actually, it's that if you don't pass a color explicitly, the default value is the string "variable" which maps to a categorical value. If you set color="something else" then you're overriding that :)

In the case of px.line() I believe that by default we set line_group to "variable" as well.

[gvwilson]

Hi - we are trying to tidy up the stale issues and PRs in Plotly's public repositories so that we can focus on things that are still important to our community. Since this one has been sitting for several years, I'm going to close it; if it is still a concern, please add a comment letting us know what recent version of our software you've checked it with so that I can reopen it and add it to our backlog. Thanks for your help - @gvwilson