CODEWITHSUNDEEP
rtransformplyr
user3857437
user3857437

Reputation: 57

Proportion by subset within columns

I think I am missing something obvious, I have the following data frame

df <- data.frame(type = c("cattle", "mixed", "not stated", "other", "sheep", "cattle", "mixed", "not stated", "other", "sheep", "cattle", "mixed", "not stated", "other", "sheep"),
        region = c("EA", "EA", "EA", "EA", "EA", "NW", "NW", "NW", "NW", "NW", "S", "S", "S", "S", "S" ),
        number = c(14, 9, 80, 0, 2, 36, 15, 45, 0, 7, 12, 35, 92, 18,  1))

I would like to calculate the proportion of type within each region. I have tried both:

require(plyr)

ddply(df, .(region, type), mutate,
prop = number/sum(number))

and

transform(df, prop = number/ave(number, region, type, FUN = sum))

Which give

         type region number prop
1      cattle     EA     14    1
2       mixed     EA      9    1
3  not stated     EA     80    1
4       other     EA      0  NaN
5       sheep     EA      2    1
6      cattle     NW     36    1
7       mixed     NW     15    1
8  not stated     NW     45    1
9       other     NW      0  NaN
10      sheep     NW      7    1
11     cattle      S     12    1
12      mixed      S     35    1
13 not stated      S     92    1
14      other      S     18    1
15      sheep      S      1    1

Thanks for reading

Upvotes: 2

Views: 81

Answers (1)

suhao399
suhao399

Reputation: 648

Actually, you need to apply ddply and group by "region" only.

Try this:

ddply(df, .(region), mutate, prop = number/sum(number))

     type region number        prop
1      cattle     EA     14 0.133333333
2       mixed     EA      9 0.085714286
3  not stated     EA     80 0.761904762
4       other     EA      0 0.000000000
5       sheep     EA      2 0.019047619
6      cattle     NW     36 0.349514563
7       mixed     NW     15 0.145631068
8  not stated     NW     45 0.436893204
9       other     NW      0 0.000000000
10      sheep     NW      7 0.067961165
11     cattle      S     12 0.075949367
12      mixed      S     35 0.221518987
13 not stated      S     92 0.582278481
14      other      S     18 0.113924051
15      sheep      S      1 0.006329114

Reason: You want to have a summary on each group by the region so you need to ddply on region only

Upvotes: 1

Related Questions