Cross Tabulation in R Dataframe - r

Question

I have a dataframe in R:
Subject T O E P Score
1 0 1 0 1 256
2 1 0 1 0 325
2 0 1 0 1 125
3 0 1 0 1 27
4 0 0 0 1 87
5 0 1 0 1 125
6 0 1 1 1 100
This is just a display of the dataframe. In reality, I have a lot of lines for each of the subjects. But the subjects are only from 1 to 6
For each Subject, the possible values are:
T : 0 or 1
O : 0 or 1
E : 0 or 1
P : 0 or 1
Score : Numeric value
I want to create a new dataframe with 6 lines (one for each subject) and the calculated MEAN score for each of these combinations :
T , O , E , P , TO , TE, TP, OE , OP , PE , TOP , TOE , POE , PET
The above will the columns of the new dataframe.
The final output should look like this
Subject T O E P TO TE TP OE OP PE TOP TOE POE PET
1
2
3
4
5
6
For each of these lines x columns the value is the MEAN SCORE
I tried aggregate and table but I can't seem to get what I want
Sorry I am new to R
Thanks

Answer 1

I had to rebuild sample data to answer the question as I understood it, tell me if it works for you :
set.seed(2)
df <- data.frame(subject=sample(1:3,9,T),
T = sample(c(0,1),9,T),
O = sample(c(0,1),9,T),
E = sample(c(0,1),9,T),
P = sample(c(0,1),9,T),
score=round(rnorm(9,10,3)))
# subject T O E P score
# 1 1 1 0 0 1 12
# 2 3 1 0 1 0 9
# 3 2 0 1 0 1 13
# 4 1 1 0 0 0 3
# 5 3 0 1 0 1 14
# 6 3 0 0 1 0 13
# 7 1 1 0 1 0 17
# 8 3 1 0 1 0 12
# 9 2 0 0 1 1 14
cols1 <- c("T","O","E","P")
df$comb <- apply(df[cols1],1,function(x) paste(names(df[cols1])[as.logical(x)],collapse=""))
# subject T O E P score comb
# 1 1 1 0 0 1 12 TP
# 2 3 1 0 1 0 9 TE
# 3 2 0 1 0 1 13 OP
# 4 1 1 0 0 0 3 T
# 5 3 0 1 0 1 14 OP
# 6 3 0 0 1 0 13 E
# 7 1 1 0 1 0 17 TE
# 8 3 1 0 1 0 12 TE
# 9 2 0 0 1 1 14 EP
library(tidyverse)
df %>%
group_by(subject,comb) %>%
summarize(score=mean(score)) %>%
spread(comb,score) %>%
ungroup
# # A tibble: 3 x 7
# subject E EP OP T TE TP
# * <int> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl>
# 1 1 NA NA NA 3 17.0 12
# 2 2 NA 14 13 NA NA NA
# 3 3 13 NA 14 NA 10.5 NA
The second step in base R:
means <- aggregate(score ~ subject + comb,df,mean)
means2 <- reshape(means,timevar="comb",idvar="subject",direction="wide")
setNames(means2,c("subject",sort(unique(df$comb))))
# subject E EP OP T TE TP
# 1 3 13 NA 14 NA 10.5 NA
# 2 2 NA 14 13 NA NA NA
# 5 1 NA NA NA 3 17.0 12

Answer 2

I'd do it like this:
# using your table data
df = read.table(text =
"Subject T O E P Score
1 0 1 0 1 256
2 1 0 1 0 325
2 0 1 0 1 125
3 0 1 0 1 27
4 0 0 0 1 87
5 0 1 0 1 125
6 0 1 1 1 100", stringsAsFactors = FALSE, header=TRUE)
# your desired column names
new_names <- c("T", "O", "E", "P", "TO", "TE", "TP", "OE",
"OP", "PE", "TOP", "TOE", "POE", "PET")
# assigning each of your scores to one of the desired column names
assign_comb <- function(dfrow) {
selection <- c("T", "O", "E", "P")[as.logical(dfrow[2:5])]
do.call(paste, as.list(c(selection, sep = "")))
}
df$comb <- apply(df, 1, assign_comb)
# aggregate all the means together
df_agg <- aggregate(df$Score ~ df$comb + df$Subject, FUN = mean)
# reshape the data to wide format
df_new <- reshape(df_agg, v.names = "df$Score", idvar = "df$Subject",
timevar = "df$comb", direction = "wide")
# clean up the column names to match your desired output
# any column names not found will be added as NA
colnames(df_new) <- gsub("df\\$|Score\\.", "", colnames(df_new))
df_new[, new_names[!new_names %in% colnames(df_new)]] <- NA
df_new <- df_new[, c("Subject", new_names)]
With the result:
> df_new
Subject T O E P TO TE TP OE OP PE TOP TOE POE PET
1 1 NA NA NA NA NA NA NA NA 256 NA NA NA NA NA
2 2 NA NA NA NA NA 325 NA NA 125 NA NA NA NA NA
4 3 NA NA NA NA NA NA NA NA 27 NA NA NA NA NA
5 4 NA NA NA 87 NA NA NA NA NA NA NA NA NA NA
6 5 NA NA NA NA NA NA NA NA 125 NA NA NA NA NA
7 6 NA NA NA NA NA NA NA NA NA NA NA NA NA NA