Question

This is actually a continuation of a previous question:

Spline on multiple factors in data frame

so apologies for going through the first part again. And believe me I tried to solve myself!

So some sample data:

mydf<- data.frame(c("a","a","b","b","c","c"),c("e","e","e","e","e","e")
                 ,as.numeric(c(1,2,3,10,20,30)),
                 as.numeric(c(5,10,20,20,15,10)))

Give some names:

colnames(mydf)<-c("Model", "Class","Seconds", "Speed")

which gives:

> mydf
  Model Class Seconds Speed
1     a     e       1     5
2     a     e       2    10
3     b     e       3    20
4     b     e      10    20
5     c     e      20    15
6     c     e      30    10

Now originally I needed to spline this data and the answer was kindly given as:

library("plyr")
ddply(mydf, .(Model), summarise, Spline = spline(x = Seconds, y = Speed))

which resulted in:

  Model                        Spline
1     a  1.0, 1.2, 1.4, 1.6, 1.8, 2.0
2     a             5, 6, 7, 8, 9, 10
3     b 3.0, 4.4, 5.8, 7.2, 8.6, 10.0
4     b        20, 20, 20, 20, 20, 20
5     c        20, 22, 24, 26, 28, 30
6     c        15, 14, 13, 12, 11, 10

Just as a side note I couldn’t even by hand create this exact structure in R so as to provide an example, even by hand!

So back to the question. I need to change the results so they look something like this:

Model   Seconds Speed
a       1.0     5
a       1.2     6
a       1.4     7
a       1.6     8
a       1.8     9
a       2.0     10
b       3.0     20
b       4.4     20
b       5.8     20
b       7.2     20
b       8.6     20
b       10.0    20
c       20      15
c       22      14
c       24      13
c       26      12
c       28      11
c       30      10

Thanks for any help!

Answer 1

You mean like this:

foo <- function(x){ 
    Spline <- spline(x = x$Seconds, y = x$Speed)
    data.frame(Seconds = Spline$x,Speed = Spline$y) 
}
out <- ddply(mydf, .(Model),foo )

Now, you might be asking yourself, “How did joran figure that out?” Cuz before I read this question, I hadn’t the foggiest notion what spline returned.

So my first stop was ?spline, and I skip to the Value section, where I find:

spline returns a list containing components x and y which give the ordinates where interpolation took place and the interpolated values.

Ah ha! So for each chunk of the data that ddply ships off to our as yet unwritten function, it needs to take that list of two components and simply convert them into a data frame, so that ddply can easily stitch things back together again.

Hence I write a function that starts out by simply fitting the spline model:

Spline <- spline(x = x$Seconds, y = x$Speed)

and then takes the results and packages it into a data frame:

data.frame(Seconds = Spline$x,Speed = Spline$y)