How to label plot with external variables

Question

Say I have the following data:

require(reshape2)
require(ggplot2)
data <- data.frame(id=seq(1,5,1), var1=c(10,3,5,7,8), label1=c(2,2,2,3,2), var2=c(12, 6, 6, 8, 6), label2=c(7,7,6,7,5))

I use melt to reshape the data to prepare it for ggplot.

data_graph <- melt(data[,c(1,2,4)], id="id")

I then plot it, with the values labeling each point:

ggplot(data=data_graph, aes(y=value, x=id, group=variable, col=variable)) +
geom_line(size=2) + geom_point() + 
geom_text(aes(label=value), size=5, hjust=-.6, vjust=1.5)

However, I actually want the label1 vector labeling the line corresponding to var1 and analogously for label2 and var2. How would I do this?

Answer 1

I think to solve your problem you need to prepare a different data.frame for ggplot2, one wiht id, variable, label and value.

# dataset for var1
df1 <- cbind(data[,c(1,3,2)], 'var1')  
colnames(df1) <- c('id', 'label', 'value', 'variable')

# dataset for var2    
df2 <- cbind(data[,c(1,5,4)], 'var2')
colnames(df2) <- c('id', 'label', 'value', 'variable')  

# putting them together
data_graph <- rbind(df1,df2)

Now your graph can be generated in this way:

gplot(data=data_graph, aes(y=value, x=id, group=variable, col=variable)) +
geom_line(size=2) + geom_point() + 
geom_text(aes(label=label), size=5, hjust=-.6, vjust=1.5)

And this is the result, with the values from the label columns as labels for your data points: