"Split" up by category in plotly

Question

I am trying to plot my scatter points in plotly, each with a specified color by category.

It works fine until I start to use "split" (this line of code shows the legend) and should, by description "split" the categorial variable up, into different categories.

This is a reproducible example:

# Load data
df_soccer <- data.frame("x" = 1:4,
                        "y" = 2:5,
                        "Name" = c("Manchester United", "FC Barcelona", "FC Porto", "Borussia Dortmund"),
                        "soc_color" = c("rgb(218, 2, 14)","rgb(167, 0, 66)", "rgb(0, 147, 51)", "rgb(0, 0, 0)"))

# Use plot_ly to plot it
df_soccer %>%
  plot_ly() %>%
  add_trace(.,
        x = ~x,
        y = ~y,
        type = "scatter",
        hoverinfo = "text",
        text = paste(df_soccer$Name),
      #  split = df_soccer$Name,
        mode = "markers",
        marker = list(color =  ~soc_color, size = 20))

What I expected is this plot:

As you can see the colours are mixed up. Borussia Dortmund should be black and ManU should be red. This is what I want, solely the legend is missing:

Does anyone has some advices to get the colours right with splitting them by category and plot them in the right colour?

Much thanks in advance!

Answer 1

By using Ben's suggestion and converting the "Name" variable into factors of the proper order the colors and the Names now match.

library(plotly)

# Load data
df_soccer <- data.frame(x = 1:4,
                        y = 2:5,
                        Name = c("Manchester United", "FC Barcelona", "FC Porto", "Borussia Dortmund"),
                        soc_color = c("rgb(218, 2, 14)","rgb(167, 0, 66)", "rgb(0, 147, 51)", "rgb(0, 0, 0)"), stringsAsFactors = FALSE)

df_soccer$Name<-factor(df_soccer$Name, levels=c("Manchester United", "FC Barcelona", "FC Porto", "Borussia Dortmund"))

plot_ly(df_soccer, name=~Name) %>%
  add_trace(x = ~x,
            y = ~y,
            type = "scatter",
            mode= "markers", 
            marker=list(color=~soc_color, size = 20))

To answer your question with added a duplicate value to the table. The easiest way for me to get it to work was to build the chart with ggplot2 and then convert it to a Plotly object.

library(ggplot2)
df_soccer <- data.frame("x" = 1:4,
                        "y" = 2:5,
                        "Name" = c("Manchester United", "FC Barcelona", "FC Porto", "Borussia Dortmund"),
                        "soc_color" = c("red","dark red", "green", "black"))
df_soccer <-rbind(df_soccer[1:2,], df_soccer[1,], df_soccer[3:4,])
df_soccer$y[3]<-3.5

g<-ggplot(df_soccer, aes(x=x, y=y, color=soc_color, name=Name)) +
  geom_point( size=5) +
  scale_colour_identity(guide = "legend", labels=(df_soccer$Name), breaks=(df_soccer$soc_color))
ggplotly(g)