PlotlyJS: customizing subplot figure titles from facet categories

Specific Domains Visualization
,
1

Hi everyone,

I am trying to adapt a piece of code from Plotly into PlotlyJS. This code allows editing the titles of individual subplots produced with facet_row and facet_col by changing the default format. I have tried hard, but somehow I failed to translate the code.

The MWE is as follows. With the code:

using PlotlyJS
using DataFrames
using CSV

df_tips =CSV.read("tips.csv" ,  delim=';' , DataFrame)
fig = plot(df_tips, x=:total_bill, y=:tip, facet_col=:sex, color=:smoker, mode="markers")

and with the tips.csv dataset (from here ), I can generate the plot below:

aa
aa1129×639 45.5 KB

I need to remove the “facet_col” name: sex=, which appears in all subplots. In this simple example, this repetition is irrelevant, but in figures with many subplots, this particular characteristic turns out to be quite disturbing.

There is a discussion of this problem here, and on the Plotly website here there is an explanation of how it can be implemented:

import plotly.express as px
fig = px.scatter(px.data.tips(), x="total_bill", y="tip", facet_col="smoker")
fig.for_each_annotation(lambda a: a.update(text=a.text.split("=")[-1]))
fig.show()

I am missing how to adapt the third code line above to the context of PlotlyJS:

fig.for_each_annotation(lambda a: a.update(text=a.text.split("=")[-1]))

Help will be very much appreciated. Thanks.

2

You can customise the label with e.g. labels=Dict(:sex => "V")

fig = plot(df_tips, x=:total_bill, y=:tip, facet_col=:sex, color=:smoker, mode="markers",
        labels=Dict(:sex => "V"))

The = remains, however, in this instance.

3

@jd-foster, that helps a little, but it looks awkward to see a useless = in all subplots. Thanks for helping.

4

Ok, try this one?:

fig = plot(df_tips, Layout(); x=:total_bill, y=:tip, facet_col=:sex, color=:smoker, mode="markers")

for sp in fig.plot.layout.annotations
        sp.text = split(sp.text,"=")[end]
end
display(fig)
5

@jd-foster, sorry, it still comes out exactly as my initial plot, in VSCode.

6

Did you get a new plot with display(fig)?

7

Yes, I did get a new plot, but it looks exactly equal to the one above.

8

I’m not sure why we have different results then. I’m using PlotlyJS v0.18.8 in VSCode.

9

I am on a Windows 10 machine, Julia 1.7.1, using PlotlyJS v0.18.8, while VSCode is 1.68.0.

I restarted VSCode to avoid some possible strange conflicts. Used exactly your code:

fig = plot(df_tips, Layout(); x=:total_bill, y=:tip, facet_col=:sex, color=:smoker, mode="markers")

    for sp in fig.plot.layout.annotations
        sp.text = split(sp.text,"=")[end]
    end

display(fig)

and my output is this:

aa
aa1124×645 38.8 KB

Thank you very much for helping.

10

It looks like you are using a pop-up window instead of the VSCode Julia Plots panel?
For some reason, please do relayout!(fig) instead of display(fig) in this case.

11

Congrats, you did it. It works by adding relayout!(fig) and calling the fig again at the end. Thanks a lot. This turnaround is extremely useful when dealing with large panel data sets.

For posterity, the code that works in VSCode, with an external plots window, is the solution provided by @jd-foster above:

fig = plot(df_tips, Layout(); x=:total_bill, y=:tip, facet_col=:sex, color=:smoker, mode="markers")

    for sp in fig.plot.layout.annotations
        sp.text = split(sp.text,"=")[end]
    end
relayout!(fig)
fig

with no more sex= in the plot:

a
a1114×630 37.7 KB