CODEWITHSUNDEEP
rgraphbar-chartfactors
LACollins
LACollins

Reputation: 13

Graphing in R - combining factors

This is a slightly odd request, I don't know if a function is necessarily likely to exist, but I was hoping there might be. Basically, I have a number of variables that I'm looking at to give an outcome and so my typical data would look like below

Sample   Daylength  Expt   Line    Protein

1            LD      L      K        100
2            SD      S      R        150
3            LD      L      R        200
4            SD      S      K        120
5            LD      L      K        95
6            SD      S      R        160
7            LD      L      R        195
8            SD      S      K        130

So I have 3 dependant variables (daylength, expt, line) and 1 outcome variable of protein. However, what I'd like to be looking at in a graph is comparing the protein levels showing bars for L and S experiments

An example of the code I would use to draw a bar graph with ggplot2 is: 

ggplot(data=results, aes(x=daylength, y=protein, fill=line)) + geom_bar(stat="identity", position=position_dodge())

and this would give me a graph showing the results for LD and SD experiments with the lines offering the separate series. However, this does not account for my third factor, the expt.

Ordinarily in Excel I would be able to manipulate the table to give results for LD/L, LD/S, SD/L and SD/S as being separate headings and create a graph using them as the X factors, however this can be arduous, especially if I'm going to rewrite the table, save it and run it in R each time. I was looking to see if there was a simple enough method to either manipulate the table in R to group everything as below with a few commands or combine the two factors to draw a graph directly each time, especially because in some cases I may be interested in combining different factors.

Sample   Daylength/Expt   Line    Protein

1            LD/L          K        100
2            SD/S          R        150
3            LD/L          R        200
4            SD/S          K        120
5            LD/L          K        95
6            SD/S          R        160
7            LD/L          R        195
8            SD/S          K        130

Upvotes: 1

Views: 398

Answers (2)

jaimedash
jaimedash

Reputation: 2743

You can use the interaction operator : on factors:

library(ggplot2)
head(CO2) # using builtin data
#  Plant   Type  Treatment conc uptake
#1   Qn1 Quebec nonchilled   95   16.0
#2   Qn1 Quebec nonchilled  175   30.4
#3   Qn1 Quebec nonchilled  250   34.8
#4   Qn1 Quebec nonchilled  350   37.2
#5   Qn1 Quebec nonchilled  500   35.3
#6   Qn1 Quebec nonchilled  675   39.2
qplot(conc, uptake, color=Type:Treatment, data=CO2)

Upvotes: 1

Pyetras
Pyetras

Reputation: 1502

You can create the Daylength/Expt column using paste:

dl_ex <- paste(results$Daylength, results$Expt, sep='/')

(Use as.character on the columns if R complains about factors or gives you weird numeric results)

and then add it to a dataset

results2 <- results[, -c("Daylength", "Expt")]
results2$DE <- dl_ex

However if you'd like to group your data by some attributes you should look into R's package plyr or sqldf if you're familiar with SQL syntax.

Upvotes: 0

Related Questions