y axis labels does not appear in scatter plot when using plotly.express with 2 inputs

[npcomplete2023]

Hello Plotly,

I experience the following problem,
when trying to generate a figure of type 'line' with multiple traces using the plotly.express
(the traces are defined implicitly, by defining a list of list in the y argument)
the resulting figure indeed includes 2 traces as expected,
it DOES name correctly the x axis but DOESN'T name correctly the y axis label, which is set to 'value' for some reason.
See attached an example code and a resulting graph (saved as screenshot)

MOREOVER, you can see the same phenomenon in the plotly tutorial
https://www.youtube.com/watch?v=GGL6U0k8WYA&t=622s
at minutes 08:50-09:00 of the video. Notice the y-axis label in the video.

Thank you

--------------------Example code--------------

import plotly.express as px
pxFig = px.line(x=[0,1,2,3],y=[[10,11,12,13],[12,13,14,15]], labels=dict(x='xLabel', y='yLabel'), title='my Title')
pxFig.show()

---

Attached image

[Coding-with-Adam]

hi @npcomplete2023
Thanks for reporting. This does look like a bug.

[Mizba-Anjum]

Hello, I tried tackling this problem in #4474 Let me know if this helps, since it's my first time contributing. The issue seemed like it was a string literal "value" in the express folder

[Coding-with-Adam]

Thank you so much @Mizba-Anjum for opening a PR and trying to tackle this issue.

[alexcjohnson]

From @nicolaskruchten in #4474:

In wide mode, the cross dimension is internally treated as if it was called value so this already works, in all orientations:

import plotly.express as px
df = px.data.medals_wide(indexed=True)
px.bar(df, labels=dict(value="hey there"))
px.bar(df, labels=dict(value="hey there"), orientation='h')

There's probably a discoverability problem but my recommendation would be to add to the docs rather than modify this somewhat complex code.

There's already some mention of this in the docs here https://plotly.com/python/wide-form/#labeling-axes-legends-and-hover-text but things could always be made clearer :)

@npcomplete2023 @Mizba-Anjum does this solve the issue for you? If so, @LiamConnors let's see what we can do to better document this. I get the feeling that there are quite a few things PX can do that I haven't discovered myself! I'll also mention this note:

labels accepts a dict whose keys are the label to rename and whose values are the desired labels.

Which is indeed exactly what you need to do here: the automatically-provided label is value so labels=dict(value='new thing') replaces it.

[Mizba-Anjum]

I can confirm that using "value" works for the y label in the original example

import plotly.express as px
pxFig = px.line(x=[0,1,2,3],y=[[10,11,12,13],[12,13,14,15]], labels=dict(x='xLabel', value='yLabel'), title='my Title')
pxFig.show()

In which case it's probably best not to change anything as it seems the actual 'y' value controls something else. I think the documentation does indeed describe it well enough though there is always room for improvement on that end. For now, I think this issue and my pull request should be closed.

[Coding-with-Adam]

@LiamConnors I'm closing this issue. Feel free to open if you think we need to update the docs.