Question

I’m analyzing USA election data, candidates contribution, etc. So I got raw data from internet and trying to learn some R executing some exercises in it. This is a CSV file that I successfully loaded and analyzed with ?summary.

I also used ?tapply successfully, to separate candidates money contribution by state:

data_amt_st = tapply(data$contb_receipt_amt, data[c('cand_nm', 'contbr_st')], sum)

?str (for a small sample) tells me the format of this data:

> str(data_amt_st)
 num [1:3, 1:21] NA NA 451 NA NA 201 NA NA 200 NA ...
 - attr(*, "dimnames")=List of 2
  ..$ cand_nm  : chr [1:3] "Bachmann, Michele" "Obama, Barack" "Romney, Mitt"
  ..$ contbr_st: chr [1:21] "33" "46" "48" "7" ...

Now I need to filter out values from data_amt_st. I need states that “Obama, Barack” had more contributions than other candidates, but don’t know how to do. Something with ?subset?

Thank you very much.

---

EDIT 1: Attending what guys told me, about making a more concrete question: I need a list of the states where Barack Obama achieved a higher contribution level (more money) than other candidates.

---

EDIT 2: Trying to give you a reproducible example (is it correct?):

x = c("a", "b", "c")
y0 = c(3, 5, 1)
y1 = c(2, 1, 6)
y2 = c(4, 2, 3)
m = cbind(x, y0, y1, y2)
m
#      x   y0  y1  y2 
# [1,] "a" "3" "2" "4"
# [2,] "b" "5" "1" "2"
# [3,] "c" "1" "6" "3"

Now, I need to know, for what y values, a is higher than b and c.

Answer 1

Maybe

## max by column (MARGIN=2)
max_amt <- apply(data_amt_st,MARGIN=2,max,na.rm=TRUE)  
data_amt_st[,max_amt==data_amt_st["Obama",]]

?
(Not sure how this will work with NA values in the Obama row: using dput to give us a reproducible example ( http://tinyurl.com/reproducible-000 ) would be useful …)

x <- letters[1:3]
y0 <- c(3, 5, 1)
y1 <- c(2, 1, 6)
y2 <- c(4, 2, 3)
m <- data.frame(y0, y1, y2)
rownames(m) <- x
maxvals <- apply(m,2,max,na.rm=TRUE)
which(m["a",]==maxvals)  ## or
names(m)[m["a",]==maxvals]