How to count levels of a factor in a data.frame, grouped by another value of that data.frame [GNU R]

Question

I have a data.frame like this

VAR1 VAR2
1999 USA
1999 USA
1999 UK
2000 GER
2000 USA
2000 GER
2000 USA
2001 USA

How do I count any level of VAR2 for each year?

What I want is a plot, where the x-axe is the year, and the y-axe is the count of any level in VAR2

And check this [question](http://stackoverflow.com/q/10130579/1478381) for another simillar reshaping problem. — Simon O'Hanlon, May 17 '13 at 08:37

score 3 · Answer 1 · answered May 17 '13 at 08:38

3

The data.table solution

library(data.table)

new.dat = data.table(dat)[,length(unique(var2)),by=var1]
new.dat=as.matrix(new.dat)
plot(x=new.dat[,1],y=new.dat[,2])

answered May 17 '13 at 08:38

Remi.b

17,389
28
87
168

score 3 · Answer 2 · answered May 17 '13 at 13:58

3

The simplest way I can think of:

let dat = your data frame

with(dat,table(VAR1,VAR2))

The output will look something like this:

      VAR2
VAR1   GER UK USA
  1999   0  1   2
  2000   2  0   2
  2001   0  0   1

Hope this helps.

answered May 17 '13 at 13:58

aaditya_n

71
2

score 1 · Accepted Answer · answered May 17 '13 at 08:32

1

There are a large number of ways and this question is undoubtedly a duplicate. What have you tried? You can use dcast in the reshape2 pacakge.

require(reshape2)
dcast( df , Country ~ Year , length )

#  Country 1999 2000 2001
#1     GER    0    2    0
#2      UK    1    0    0
#3     USA    2    2    1

answered May 17 '13 at 08:32

Simon O'Hanlon

58,647
14
142
184

to plot this data with ggplot, this code works for me: ggplot(df, aes(x=factor(Year),fill=Country)) + geom_bar(position="dodge") – Produnis May 24 '13 at 06:54

How to count levels of a factor in a data.frame, grouped by another value of that data.frame [GNU R]

3 Answers3