Change colour in 2 different ggplot stat_summary functions

Question

My question is an extension of this one but I have used 2 stat_summary calls in my ggplot. The first stat_summary shows the mean of each unit and the second stat_summary shows the overall mean for a group (time) of multiple units (see attached plot). Assigning the colours of the first stat_summary is no problem using colour=factor(unit) and specifying with scale_colour_manual(). I would like to assign colours in the second stat_summary by the 2 levels of time (pre and pst) e.g. colour=time, but i cannot then use scale_colour_manual again to specify colours. I have tried: colour=c("red", "blue"), colour=cbPalette2, aes(color=paste("mean", time))) (as suggested in the related question), but no luck. Any help would be much appreciated.

data:

 structure(list(time = c("pre", "pre", "pre", "pre", "pre", "pre","pre", "pre", "pre", "pre", "pre", "pre", "pre", "pre", "pre", "pre", "pre", "pre", "pre", "pre", "pre", "pre", "pre", "pre", "pst", "pst", "pst", "pst", "pst", "pst", "pst", "pst", "pst", "pst", "pst", "pst", "pst", "pst", "pst", "pst", "pst", "pst", "pst", "pst", "pst", "pst", "pre", "pre", "pre", "pre", "pre", "pre", "pre", "pre", "pre", "pre", "pre", "pre", "pre", "pre", "pre", "pre", "pre", "pre", "pre", "pre", "pre", "pst", "pst", "pst", "pst", "pst", "pst", "pst", "pst", "pst", "pst", "pst", "pst", "pst", "pst", "pst", "pst", "pst", "pst", "pst", "pst", "pst", "pst", "pst", "pst", "pre", "pre", "pre", "pre", "pre", "pre", "pre", "pre", "pre", "pre", "pre", "pre", "pre", "pre", "pre", "pre", "pre", "pre", "pre", "pre", "pst", "pst", "pst", "pst", "pst", "pst", "pst", "pst", "pst", "pst", "pst", "pst", "pst", "pst", "pst", "pst", "pst", "pst"), unit = c(73, 72, 72, 73, 73, 75, 75, 75, 72, 72, 73,73, 75, 75, 72, 74, 71, 74, 74, 71, 71, 74, 74, 71, 72, 73, 72, 72, 73, 73, 72, 73, 72, 75, 75, 75, 75, 75, 71, 71, 71, 74, 74, 71, 74, 71, 70, 66, 68, 66, 68,70, 70, 66, 70, 68, 66, 68, 70, 67, 69, 69, 69, 67, 69, 67, 69, 66, 66, 66, 68, 68, 66, 68, 66, 68, 70, 70, 70, 70, 70, 67, 67, 69, 67, 69, 69, 67, 69, 67, 69, 76, 76, 76, 78, 78, 78, 76, 78, 78, 76, 77, 77, 77, 79, 79, 79, 77, 77, 79, 79, 78, 78, 78, 78, 76, 76, 76, 76, 76, 77, 79, 79, 77, 77, 79, 79, 77, 79), number = c(159, 152, 151, 157, 156, 169, 167, 166, 153, 155, 158, 160, 168, 170, 154, 164, 146, 162, 161, 150, 147, 165, 163, 149, 231, 236, 235, 233, 240, 238, 232, 237, 234, 246, 247, 248, 249, 250, 226, 230, 228, 245, 243, 227, 242, 229, 141, 123, 133, 122, 132, 143, 145, 121, 142, 131, 124, 134, 144, 130, 138, 140, 137, 126, 136, 129, 139, 201, 205, 203, 215, 213, 202, 212, 204, 214, 221, 222, 223, 224, 225, 206, 210, 216, 208, 220, 218, 207, 217, 209, 219, 174, 172, 171, 182, 184, 181, 173, 185, 183, 175, 177, 179, 176, 189, 186, 187, 180, 178, 190, 188, 262, 261, 265, 263, 251, 252, 253, 254, 255, 257, 269, 267, 256, 260, 266, 270, 258, 268), rep = c("A", "A", "A", "A", "A", "A", "A", "A", "A", "A", "A", "A", "A", "A", "A", "A", "A", "A", "A", "A", "A", "A", "A", "A", "A", "A", "A", "A", "A", "A", "A", "A", "A", "A", "A", "A", "A", "A", "A", "A", "A", "A", "A", "A", "A", "A", "B", "B", "B", "B", "B", "B", "B", "B", "B", "B", "B", "B", "B", "B", "B", "B", "B", "B", "B", "B", "B", "B", "B", "B", "B", "B", "B", "B", "B", "B", "B", "B", "B", "B", "B", "B", "B", "B", "B", "B", "B", "B", "B", "B", "B", "C", "C", "C", "C", "C", "C", "C", "C", "C", "C", "C", "C", "C", "C", "C", "C", "C", "C", "C", "C", "C", "C", "C", "C", "C", "C", "C", "C", "C", "C", "C", "C", "C", "C", "C", "C", "C", "C"), tied = c("C", "C", "C", "C", "C", "C", "C", "C", "C", "C", "C", "C", "C", "C", "C", "V", "V", "V", "V", "V", "V", "V", "V", "V", "C", "C", "C", "C", "C", "C", "C", "C", "C", "C", "C", "C", "C", "C", "V", "V", "V", "V", "V", "V", "V", "V", "C", "C", "C", "C", "C", "C", "C", "C", "C", "C", "C", "C", "C", "V", "V", "V", "V", "V", "V", "V", "V", "C", "C", "C", "C", "C", "C", "C", "C", "C", "C", "C", "C", "C", "C", "V", "V", "V", "V", "V", "V", "V", "V", "V", "V", "C", "C", "C", "C", "C", "C", "C", "C", "C", "C", "V", "V", "V", "V", "V", "V", "V", "V", "V", "V", "C", "C", "C", "C", "C", "C", "C", "C", "C", "V", "V", "V", "V", "V", "V", "V", "V", "V"), measure = c(0.65, 0.32, 0.29, 0.53, 1.18, 0.62, 0.2, 0.38, 0.64, 0.29,0.45, 0.66, 0.18, 0.74, 0.36, 0.97, 0.27, 0.2, 0.5, 0.34, 0.19, 0.73, 0.58, 0.29, 0.47, 0.17, 0.75, 0.47, 0.37, 0.44, 0.79, 0.37, 0.51, 0.19, 0.28, 0.21, 0.18, 0.2, 0.01, 0.69, 0.66, 0.32, -0.05, 0.69, -0.0899999999999999, 0.52, 0.2, 0.17, 0.21, 0.2, 0.17, 0.31, 0.41, 0.38, 0.21, 0.47, 0.37, 0.29, 0.52, 0.26, 0.47, 0.27, 0.35, 1.14, 1.23, 0.81, 0.56, 0.64, 0.35, 0.43, 0.51, 0.44, 0.84, 0.59, 0.14, 0.55, 0.16, 0.13, 0.16, 0.15, 0.17, 0.31, 0.6, 0.22, 0.57, 0.65, 0.51, 0.77, 0.56, 0.51, 0.52, 0.2, 0.18, 0.03, 0.12, 0.37, 0.14, 0.1, 0.31, 0.16, 0.39, 0.2, 0.27, 0.17, 0.27, 0.14, 0.13, 0.28, 0.5, 0.23, 0.17, 0.05, 0.67, 0.4, 0.24, 0.18, 0.33, 0.22, 0.12, 0.15, 0.0700000000000001, 0, 0.02, 0.39, 0.41, 0.88, 0.28, -0.0600000000000001, -0.01)), .Names = c("time", "unit", "number", "rep", "tied", "measure"), row.names = c(NA,-129L), class = "data.frame")

code:

 cbPalette2 <- c("ghostwhite", "lightsteelblue1", "lightyellow" )
 unitcolours1<-c("purple","gold", "#999999","#009E73","plum1","#E69F00","#56B4E9","black","lightgreen","red","#0072B2","darkolivegreen3","#CC79A7","deeppink4")


 v<-ggplot(stack, aes(x=tied, y=measure, fill=time)) + 

 geom_point(size=0.5, aes(fill=factor(time), colour=factor(unit)), position=position_jitterdodge(jitter.width = NULL, jitter.height = 0, dodge.width = 0.75))+

    stat_summary(mapping=aes(shape=95, fill=time, colour=factor(unit)), fun.y= "mean", geom="point", size=6, position=position_dodge(width=0.75))+

    stat_summary(mapping=aes(shape=95, fill=time) , fun.y= "mean", geom="point", size=16, position=position_dodge(width=0.75))+

 facet_grid(.~rep)+
 labs(fill="time",  x="tied", y="measure (cP)")+
 scale_fill_manual(values=cbPalette2) +
 scale_colour_manual(values=unitcolours1)+
 guides(colour=FALSE)+
 scale_shape_identity() +
 theme_bw()+
 scale_y_continuous(limits=c(-0.1,1.24), breaks=seq(0, 1.2, 0.1))

I'm a bit confused about what you're trying to do. You have `fill=time` in the main call to `ggplot`, but none of your geoms (which are all `geom_point`) take a fill aesthetic. Then you repeat `fill=time` within each `stat_summary` and you have `fill=factor(time)` within `geom_point`. These are unnecessary, if you have fill in the main call to ggplot, and the marker shapes don't take a fill aesthetic anyway. Also, you have `shape=95` inside `aes`. Unless you're mapping shape to a column in your data, `shape=95` should be outside `aes`. — eipi10, Nov 01 '16 at 19:36

score 1 · Answer 1 · answered Nov 01 '16 at 19:41

1

As I noted in my comment, I'm not sure how you want the plot to look, but you can get separate color mappings by using a filled marker (such as shapes 21 - 25) for one geom and a non-filled marker for another. Then you can map a variable to the fill aesthetic for the filled marker and map another variable to the colour aesthetic for the non-filled marker.

Here's an example, using a modified version of your code. This probably isn't quite what you had in mind, so let me know what you were actually trying to do and I'll modify as necessary.

pd = position_dodge(0.75)

ggplot(stack, aes(x=tied, y=measure)) + 
  geom_point(size=1, aes(fill=time), shape=21, colour=NA, 
             position=position_jitterdodge(jitter.width=NULL, jitter.height=0, dodge.width=0.75)) +
  stat_summary(shape=95, fun.y=mean, geom="point", size=12, position=pd) +
  stat_summary(shape=95, aes(colour=factor(unit)), fun.y=mean, geom="point", size=8, position=pd) +
  facet_grid(.~rep) +
  labs(fill="Time", x="tied", y="measure (cP)") +
  # scale_fill_manual(values=cbPalette2) +
  # scale_colour_manual(values=unitcolours1) +
  guides(colour=FALSE) +
  #scale_shape_identity() +
  theme_bw() +
  scale_y_continuous(limits=c(-0.1,1.24), breaks=seq(0, 1.2, 0.1))

answered Nov 01 '16 at 19:41

eipi10

91,525
24
209
285

Hi, thank you for your efforts. I keep trying to upload a picture would explain more clearly what i mean, but it's not working for some reason. The graph i am trying to produce does of course look a lot like the one you have produced, but there should not be 6 black lines (the ones i am trying to colour), there should be 12 (one for each time (pre, pst), and within each time (pre, pst) there should also be double the coloured lines; this is where the fill=factor(time) comes in, i put it in each geom to subgroup each tied, without it all the dots just group together. – Rebecca Nov 01 '16 at 20:46
Using `colour=time` will give different colors to the two levels of `time`. However, `factor(time)` does exactly the same thing as `time`, whether you use it with `fill` or with `colour`. You need to map a *different* data column to `fill` to get a second set of colors within each level of `time`. Which column do you want to use for the `fill` values? – eipi10 Nov 04 '16 at 16:37
Hi, the dots should be coloured according to unit (which i can do with scale_colour_manual, see original code) and the lines should be coloured according to time. Fill is currently being used to seperate the dots by time (a sub group of tied), but doesn't seem to want to colour the black lines accordingly. I could use group instead of fill which still colours the dots correctly, but doesn't seperate them properly according to time. – Rebecca Nov 07 '16 at 13:07

Change colour in 2 different ggplot stat_summary functions

1 Answers1

Linked