I would begin by visualizing my data to get some intuitions , detect patterns and choose the best model. R is really great for this. Here I am using ggplot2
but you can do the same thing using lattice
. For example :
dat$DATE <- as.Date(dat$DATE,format="%d/%m/%Y")
library(ggplot2)
ggplot(dat,aes(x=DATE,y=Rev,group=ProdCode,color=ProdCode)) +
geom_line() +
facet_grid(~ID) +
geom_smooth(method='lm')
Here dat are :
dat <- read.table(text="DATE ID ProdCode Rev
08/01/2012 1-100FDK B51 0.22
09/01/2012 1-100FDK B51 10.17
10/01/2012 1-100FDK B51 39.73
11/01/2012 1-100FDK B51 78.77
12/01/2012 1-100FDK B51 167.66
08/01/2012 1-100FDK B53 387.82
09/01/2012 1-100FDK B53 321.58
10/01/2012 1-100FDK B53 261.9
11/01/2012 1-100FDK B53 259.84
12/01/2012 1-100FDK B53 269.84
08/01/2012 1-200FFK A54 194.42
09/01/2012 1-200FFK A54 225.1
10/01/2012 1-200FFK A54 269.42
11/01/2012 1-200FFK A54 321.14
12/01/2012 1-200FFK A54 429.11",
header=TRUE)