Change the line style depending on the third variable - R

Question

Following my data framework using dput ():

structure(list(year = 1984:2007, ger.90.10.ratio = c(3.1, 3.09, 2.98, 2.93, 2.98, 2.96, 2.95, 3.12, 3.16, 3.13, 3.24, 3.3, 3.28, 3.24, 3.24, 3.17, 3.23, 3.31, 3.63, 3.69, 3.75, 3.91, 4.1, 3.99 ), kanzler = c(1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 1L, 1L), wahljahr = c(0L, 0L, 1L, 0L, 0L, 0L, 1L, 0L, 0L, 0L, 1L, 0L, 0L, 0L, 1L, 0L, 0L, 0L, 1L, 0L, 0L, 1L, 0L, 0L), us.90.10.ratio = c(9.55, 9.69, 10.09, 10.23, 10.21, 9.99, 10.12, 10.22, 10.34, 10.64, 10.57, 10.11, 10.33, 10.6, 10.44, 10.42, 10.58, 10.63, 10.75, 11.22, 11.08, 11.17, 11.08, 11.18), us.80.20.ratio = c(4.36, 4.38, 4.49, 4.48, 4.45, 4.44, 4.42, 4.51, 4.6, 4.65, 4.68, 4.52, 4.61, 4.64, 4.65, 4.62, 4.56, 4.65, 4.69, 4.83, 4.76, 4.78, 4.84, 4.93), ger.ratio.diff = c(NA, -0.0100000000000002, -0.11, -0.0499999999999998, 0.0499999999999998, -0.02, -0.00999999999999979, 0.17, 0.04, -0.0300000000000002, 0.11, 0.0599999999999996, -0.02, -0.0399999999999996, 0, -0.0700000000000003, 0.0600000000000001, 0.0800000000000001, 0.32, 0.0600000000000001, 0.0600000000000001, 0.16, 0.19, -0.109999999999999), ger.ratio.change = c(0, -0.0100000000000002, -0.11, -0.0499999999999998, 0.0499999999999998, -0.02, -0.00999999999999979, 0.17, 0.04, -0.0300000000000002, 0.11, 0.0599999999999996, -0.02, -0.0399999999999996, 0, -0.0700000000000003, 0.0600000000000001, 0.0800000000000001, 0.32, 0.0600000000000001, 0.0600000000000001, 0.16, 0.19, -0.109999999999999)), row.names = c(NA, -24L), .Names = c("year", "ger.90.10.ratio", "kanzler", "wahljahr", "us.90.10.ratio", "us.80.20.ratio", "ger.ratio.diff", "ger.ratio.change" ), class = "data.frame") 

What I want to do is create a line graph that changes the line style depending on the value of the third variable. It should look like this

In a word: In the above data frame, I would like to build ger.90.10.ratio against the year and change the line style depending on the values ​​in the chancellor's column.

I searched everywhere (maybe my google-fu is bad?), But I really would like to know if this is possible in R.

Thank you for your help, very grateful!

Answer 1

R basic graphics do not allow you to change the properties of the line (type / dash, color, etc.) in the middle of the line, so you need to separate the data and use lines() on parts or segments() .

This is not a complete solution, but the beginning:

Set graphical parameters (unnecessary, but I like them):

 par(las=1,bty="l") 

Create a variable for splitting (we want each segment to be differentiated):

 dat <- transform(dat,k2=c(0,cumsum(diff(kanzler)!=0))) 

Share the data:

 s <- split(dat,dat$k2) 

Set up an empty plot:

 plot(ger.90.10.ratio~year,data=dat,type="n") 

Add lines for each part:

 invisible(lapply(s, function(x) with(x,lines(year,ger.90.10.ratio,col=kanzler)))) 

You can make this a bit more compact (as usual) in ggplot2 , but the k2 variable we set is still useful:

 library("ggplot2") qplot(year,ger.90.10.ratio,data=dat,colour=factor(kanzler), group=k2,geom="line") 

Answer 2

This is a solution using the lattice package to show how similar it is to basic graphics. I mainly use @Ben's idea of ​​sharing data. I use the lattice panel to give more compact .

 library(lattice) dat <- transform(dat,k2=c(0,cumsum(diff(kanzler)!=0))) s <- split(dat,dat$k2) xyplot(ger.90.10.ratio~year,data=dat,type=c('p'),groups=kanzler, main = 'THE PARTISAL POLITICAL ECONOMY', auto.key=list(columns=2,cex=2,text=c('Democrats','Republicans')), panel=function(x,y,...) { panel.xyplot(x,y,...) lapply(s,function(x) with(x,panel.lines(year,ger.90.10.ratio,col=kanzler))) panel.grid() })