ggvis color line graphs and corresponding legend

Question

I am plotting a ggvis graph with different colors and linetype by setting stroke and strokeDash parameters. The colors are not showing what I want, for ex: I need "red", "blue" and "green" colors where as the colors shown are " blue", "orange" and "green".(code example and screenshot are below).

I also need to add legend based on the color and linetype. I am unable to add the legend combining both "color type"(stroke) and "line type"(strokeDash).

Below is a sample code I am working.

mydf <- data.frame(xval = c(1:5),USA = c(11:15), Canada = c(21:25), UK = c(31:35))

mydf2 <- tidyr::gather(mydf, 'var', 'val', -xval)

mydf2$mycol <- c(rep("Red",5), rep("Blue",5), rep("Green",5))
mydf2$mydash <- c(rep(5,5), rep(-1,5), rep(10,5))

myggv2 <- mydf2 %>% ggvis(x = ~xval, y = ~val, stroke = ~mycol,      
strokeDash := ~mydash) %>% layer_points(size := 100) %>%
add_tooltip(function(d) { paste0("x:", d$xval, "<br>", "y:", d$val) }, "hover") %>% group_by(var) %>% 
layer_paths() %>% add_axis("x", title = "qtrs") %>% 
  add_axis("y", title = "MEV")%>% hide_legend("stroke")
myggv2

Here is the screenshot of the graph.Screenshot

Thanks

Answer 1

You can change colors in your plot and legend for discrete variables using scale_nominal.

In this approach, you map the variable you want in the legend within the plotting code and then change away from the default colors using your chosen colors. I'm using var for the stroke variable in this example.

mydf2 %>%
    ggvis(x = ~xval, y = ~val, stroke = ~var,      
          strokeDash := ~mydash) %>% 
    layer_points(size := 100) %>%
    add_tooltip(function(d) { paste0("x:", d$xval, "<br>", "y:", d$val) }, "hover") %>% 
    group_by(var) %>% 
    layer_paths() %>% 
    add_axis("x", title = "qtrs") %>% 
    add_axis("y", title = "MEV") %>%
    scale_nominal("stroke", label = "Country", range = unique(mydf2$mycol))

Note that at this time ggvis does not support a strokeDash legend. See this open github issue.