CODEWITHSUNDEEP
raggregate
shrimp32
shrimp32

Reputation: 147

How to calculate the mean in a data frame using aggregate function in R?

I have a data frame df1:

number=c(4,3,2,3,4,1)
year=c("2000","2000","2000", "2015", "2015", "2015")
items=c(12, 10, 15, 5, 10, 7)
df1=data.frame(number, year, items)
setDT(df1)[, Prop := number/sum(number), by = year]

such that it looks like this:

  number year items      Prop
1:      4 2000    12 0.4444444
2:      3 2000    10 0.3333333
3:      2 2000    15 0.2222222
4:      3 2015     5 0.3750000
5:      4 2015    10 0.5000000
6:      1 2015     7 0.1250000

I want to get the mean of the number of items per year, so I tried using this fuction:

mean.df1=aggregate((df1$number*df1$Prop),list(df1$year), mean)

but it returns the wrong values for the mean. I want it to return:

  Group.1        x
1    2000 2.918918
2    2015 2.296296

where Group.1 is the year and x is the correct mean.

Thanks!

Upvotes: 5

Views: 23174

Answers (3)

shrimp32
shrimp32

Reputation: 147

I just had to switch "mean" to "sum" in my aggregate function such that it becomes:

mean.df1=aggregate((df1$number*df1$Prop),list(df1$year), sum)

Upvotes: 2

Satya
Satya

Reputation: 1800

How about using the dplyr package 

library(dplyr)
df1  %>% group_by(year) %>% summarise(mean = sum(number * items)/sum(items))

which gives

  year     mean
1 2000 2.918919
2 2015 2.818182

Upvotes: 3

Rorschach
Rorschach

Reputation: 32416

To aggregate mean number of items/year

aggregate(number ~ year, data=df1, mean)
#   year   number
# 1 2000 3.000000
# 2 2015 2.666667

Edit

For the weighted average in base R you could do standard split-apply-combine

sapply(split(df1, df1$year), function(x) weighted.mean(x$number, w=x$items))

or

sapply(split(df1, df1$year), function(x) sum(x$number*x$items)/sum(x$items))
#     2000     2015 
# 2.918919 2.818182

Upvotes: 8

Related Questions