Overlay with merged colors/labels

hvPlot
1

Hi!

I am just taking my first steps with hvplots since I am working with polars which defers to it. I am creating two plots, which represent measured and ground truth data. Every plot is a set of lines for a group of devices:

scat1 = data.plot.scatter(x="x", y="measured", c="category", groupby="device", marker="x")
scat2 = data.plot.scatter(x="x", y="truth", c="category", groupby="device", marker="o")
comb = scat1*scat2

This is almost perfect, but I need the dots of the same color to also be joined by a line.

I have two issues here:

  • I couldnt find a way to combine both lines and markers. Prior solutions seems to combine a scatter and lineplot, but I can’t use the c="category" attribute on lineplots, so the resulting colors are very irritating
  • Using just the lineplot I cant even properly overlay them. The overlay has new colors assigned and the legend only shows those of one of the plots, see here:

Screenshot 2024-08-02 at 12.44.04
Screenshot 2024-08-02 at 12.44.041788×742 28.6 KB

Is there any way to overlay both plots as curves while still allowing to distinguish them (e.g. by different markers?). I tried going through the options but couldn’t find anything that would help.

2

There’s a hvplot.path I think that potentially works with c?

Other than that, looping through groups and manually multiplying probably would work.

overlay = hv.Overlay()
for group, group_df in group_df.groupby("category"):
    overlay *= group_df.hvplot(...)
3

I’ll have to try the path one

Looping through groups manually does not work. That is exactly what I did to produce the image I linked where, as you can see, even though I have twice the same categories, the lines from the plots are assigned different colors. The blue line in that picture should be green, the yellow line should be red