Subset or Grep for selectively pasting using Geom_Text in R

Question

I am trying to add select text using geom_text for a series of ggplots I am generating using the lapply and split functions. Essentially I am creating one ggplot per Component.name.x with each of these plots containing 3 curves - one for each group.name.

So as a shortened example of what my data frame (working2.df) kind of looks like:

group.name    Component.name.x         TCAmount  Ratio   Slope   Intercept
1A-Alprazolam          Alprazolam         0.0      0.05    0.5       0.4
1A-Alprazolam          Alprazolam         1.5      0.07    0.5       0.4
1A-Alprazolam          Alprazolam         3.1      0.08    0.5       0.4
1A-Alprazolam          Alprazolam         6.25     0.10    0.5       0.4
2A-Alprazolam          Alprazolam         0.0      0.06    0.4       0.3
2A-Alprazolam          Alprazolam         1.5      0.08    0.4       0.3
2A-Alprazolam          Alprazolam         3.1      0.09    0.4       0.3
2A-Alprazolam          Alprazolam         6.25     0.10    0.4       0.3
1B-Alprazolam          Alprazolam         0.0      0.05    0.5       0.6
1B-Alprazolam          Alprazolam         1.5      0.08    0.5       0.6
1B-Alprazolam          Alprazolam         3.1      0.10    0.5       0.6
1B-Alprazolam          Alprazolam         6.25     0.11    0.5       0.6
1A-Codeine             Codeine            0.0      0.10    0.5       0.6
1A-Codeine             Codeine            1.5      0.30    0.5       0.6
1A-Codeine             Codeine            3.1      0.41    0.5       0.6
1A-Codeine             Codeine            6.25     0.62    0.5       0.6
2A-Codeine             Codeine            0.0      0.09    0.6       0.7
2A-Codeine             Codeine            1.5      0.29    0.6       0.7
2A-Codeine             Codeine            3.1      0.40    0.6       0.7
2A-Codeine             Codeine            6.25     0.62    0.6       0.7
1B-Codeine             Codeine            0.0      0.09    0.6       0.5
1B-Codeine             Codeine            1.5      0.28    0.6       0.5
1B-Codeine             Codeine            3.1      0.41    0.6       0.5
1B-Codeine             Codeine            6.25     0.61    0.6       0.5

Currently I am plotting the data using the following:

plotIntra<-function(working2.df){
    ggplot(working2.df,aes(x=TCAmount,y=Ratio, group=group.name, colour=group.name))+
        geom_smooth(method=lm, level=0.95, se=FALSE, aes(colour=group.name))+
        theme_bw()+
        ggtitle(working2.df$Component.name.x[1])+
        xlab("Amount (ng/mL)") +ylab("Response Ratio")+
        expand_limits(y=0)+
        geom_text(data=working2.df, aes(label=paste("y=", Slope[1], "x+", Intercept[1], sep="")))
}

lapply(split(working2.df, working2.df$Component.name.x), function(x) try(plotIntra(x)))

What I want is for R to grab each Slope and Intercept once per group.name and paste it on the graph with the corresponding curve. So far I have only managed to end up with the y=mx+b for ALL the curves on every plot (sigh) or to have the y=mx+b text pasted many many repeating times on the same plot.

What I have been trying to do (but don't know if it is possible) is to make R either think there is a subset of data which contains the Slope and Intercept only once and it matches by group.name, or to use grep or subset in some way to pull the data. Maybe something like this (but this doesn't work):

geom_text(data=working2.df, grep("1A", working2.df$group.name)), 
    aes(label=paste("y=", Slope[1], "x+", Intercept[1], sep=""))

Anyone think they can help with this and be a ggplot hero?!

score 1 · Accepted Answer · answered Dec 09 '15 at 23:01

1

You can collapse the values in your function and use a different data= parameter for the geom_text

plotIntra<-function(working2.df){
    lbls <- aggregate(cbind(TCAmount,Ratio)~group.name+Slope+Intercept, working2.df, mean)
    ggplot(working2.df,aes(x=TCAmount,y=Ratio, group=group.name, colour=group.name))+
        geom_smooth(method=lm, level=0.95, se=FALSE, aes(colour=group.name))+
        theme_bw()+
        ggtitle(working2.df$Component.name.x[1])+
        xlab("Amount (ng/mL)") +ylab("Response Ratio")+
        expand_limits(y=0)+
        geom_text(data=lbls, aes(label=paste("y=", Slope[1], "x+", Intercept[1], sep="")))
}

lapply(split(working2.df, working2.df$Component.name.x), function(x) try(plotIntra(x)))

answered Dec 09 '15 at 23:01

MrFlick

195,160
17
277
295

That works great. It does add the letter "a" over the bullet point in the legend? Can this be removed? I also know there are instructions for moving geom_text around on the plot but would you know how to move the lines that have been pasted since I'm not sure what the "object name" of the individual text is when you use aggregate and then paste it? Thanks! – S Wellman Dec 10 '15 at 14:30
Found out from another post (http://stackoverflow.com/questions/2409357/how-to-nicely-annotate-a-ggplot2-manual) that show_guide=F added to geom_text removes the text from the bullet point. Still looking for help to move the text around the graph! – S Wellman Dec 10 '15 at 17:33
I noticed now that "mean" actually computes the mean slope and intercept of the subset generated by aggregate - so that it plots the same y=mx+b equation for each line. I want to show the y=mx+b equation of each line as it is given in the table - not the average. Can another function be used to loop through the slope and x-ints to plot? – S Wellman Dec 11 '15 at 16:29

Subset or Grep for selectively pasting using Geom_Text in R

1 Answers1