-1

I need to make a graph from a time series which has in the first column a group variable with 2 levels, in the second another group variable (smaller group) with 22 levels, followed by 427 variables that code for days with a numeric outcome. Something like this:

VT  Medication V1 V2 V3 V4 V5 V6 V7 V8 
1           1   0  1  1  1  3  3  3  3 
1           2   0  1  1  1  2  1  1  1 
1           3   0  2  2  2  0  0  0  0 
1           4   1  1  4  5  6  4  0  0
2           1   1  1  3  3  3  3  3  3
2           2  etc
2           3
2           4

now I want to make two plots, one with from the VT group 1, and the second from VT group 2 with the numeric outcome on the y-axis and the days on the x-axis. I also want it to be a cumulative graph, distinguishing the contribution to the graph by color. I was thinking that it would look almost the same as doing stacked columns per day and putting them next to one another. however if possible i would want it to be smooth lines, like stacking lines of an interaction plot on top of another.

it would have to look like a graph with 22 layers with different colors coding for the 22 "medication" -groups.

I hope it's somewhat clear what i want from this description. Dirk

Dirk van Dooren
  • 89
  • 8

1 Answers1

0

I usually convert my data to long format when I do plots like this. The ggplot2 package is a good choice to produce grouped plots of various kind.

library(reshape)
library(ggplot2)

data <- read.table(header=TRUE, text=
"VT  Medication V1 V2 V3 V4 V5 V6 V7 V8
1           1   0  1  1  1  3  3  3  3
1           2   0  1  1  1  2  1  1  1
1           3   0  2  2  2  0  0  0  0
1           4   1  1  4  5  6  4  0  0
2           1   1  1  3  3  3  3  3  3")

data <- melt(data, id.vars=c("VT","Medication"), variable_name="Day")
data$VT <- as.factor(data$VT)
data$Medication <- as.factor(data$Medication)
data$Day <- as.numeric(data$Day)
plot.data <- data[data$VT==1,]

ggplot(plot.data, aes(,x=Day, y=value, colour=Medication)) +
  geom_line() +
  scale_colour_manual(values=c("red","blue","green","black"))

enter image description here

SimonG
  • 4,701
  • 3
  • 20
  • 31
  • ok, this is somewhat what I am looking for, but how can i stack these lines on top of another? this is important to me because i will place it next to a plot called a "lasagna plot" (http://www.smart-stats.org/content/biosignals) which looks similar in structure but is built up in a different manner. – Dirk van Dooren Sep 01 '14 at 07:20
  • What do you mean exactly by "stacking them on top of another"? – SimonG Sep 01 '14 at 08:21
  • if the value on day 2 for medication1=1 and for medication 2= 2 i would want it to show in the graph that medication1=1 and medication2= 2+1 so basically medication2=medication2+medication1 (so Ni = Ni - (Ni-1)) – Dirk van Dooren Sep 01 '14 at 09:16
  • cumulative values is the word I was looking for – Dirk van Dooren Sep 01 '14 at 09:31
  • If you want it to be like that, then you should re-calculate "Medication2" accordingly before plotting. – SimonG Sep 01 '14 at 09:46