CODEWITHSUNDEEP
rplotggplot2alpha
James
James

Reputation: 1234

Set the alpha of points in ggplot to a numeric vector

I am trying to create an NMDS plot using ggplot. In addition to the standard plotting of points (communities/assemblages) and hulls etc. I would also like to adjust the alpha of each point.

I have calculated a vector of numeric values from 0-1 (called "Alpha") which reflects the timing of the surveys which correspond to each point and I would like to use this to modify the alpha of each point. Thus far I have been using variations on the theme of:

 ggplot() +
     geom_point(data = data.scores, aes(x = NMDS1, y = NMDS2, colour =Treatment), size = 3) +
     scale_colour_manual(values = c("Burnt" = "black", "Grazed" = "tan4","Control" = "green4")) +
     scale_alpha_manual(values = Alpha) +
     geom_polygon(data = hull.data, aes( x = NMDS1, y = NMDS2, colour = Treatment),
                  linetype = "dashed", fill = NA) +
     coord_equal() +
     theme_bw() +
     theme(panel.border = element_blank(), panel.grid.major = element_blank(),
           panel.grid.minor = element_blank(), axis.line = element_line(colour = "black"),
           legend.position="top")

This has only got me plots where the alpha of all the points is 1, the alpha is not 1 but is set to the same value (presumably the first value in my vector) or error messages. Does anyone have any idea what I'm doing wrong?

For reference I was able to create what I want using base R graphics fairly easily but I need my plot to be a ggplot graphic:

enter image description here

Upvotes: 1

Views: 1535

Answers (1)

Nate
Nate

Reputation: 10671

ggplot() syntax expects the values being mapped for a layer to be passed in that layer, so in geom_point(), and not by adding a scale. So right now you are not defining which layers you want the alpha scale to be applied to, thus all are being shown as alpha = 1 (the default).

You can just use the raw vector directly in geom_point(). Using a reproducible example with mtcars:

library(ggplot2)
set.seed(1)

# sim your Alpha
Alpha <- runif(nrow(mtcars))

ggplot(data = mtcars, aes(mpg, hp)) +
    geom_point(alpha = Alpha)

enter image description here

scale_alpha_manual() won't have any effect on this plot since you are not mapping anything with aes(alpha = ...).

Upvotes: 3

Related Questions