CODEWITHSUNDEEP
rdplyr
James White
James White

Reputation: 815

Counting number of values great than a percentile value across multiple factor levels

I have a data frame with a set of values and a factor. 

df <- as.data.frame(matrix(sample(0:10, 1*30, replace=TRUE), ncol=1))
colnames(df)[1] <- "values"
df$factor<- rep(c("Factor.A","Factor.B"), each = 15)

What I would like to do is calculate the 75th percentile of values within each group... 

Percentile_75 <- aggregate(values ~ factor, function(x)
quantile(x,(0.75)), data = df)

...and see how many values are greater than each of these thresholds within df for each factor level. I can do this manually for each factor, but in reality I have far more factor levels, so I am guessing there is a neat (possibly dplyr) function that would be able to do this easily? Thank you in advance.

Upvotes: 0

Views: 943

Answers (2)

Balter
Balter

Reputation: 1095

A data.table approach:

   df <- as.data.frame(matrix(sample(0:10, 1*30, replace=TRUE), ncol=1))
   colnames(df)[1] <- "values"
   df$factor<- rep(c("Factor.A","Factor.B"), each = 15)

   library(data.table)
   df <- setDT(df)

   df[,P_75 := quantile(values, probs = 0.75), by = factor][
  values > P_75, .(unique(P_75),.N), by = factor
   ]

#      factor  V1 N
# 1: Factor.A 7.5 4
# 2: Factor.B 8.0 2

Upvotes: 1

moodymudskipper
moodymudskipper

Reputation: 47330

with dplyr you can do this:

library(dplyr)
df %>%
  group_by(factor) %>% 
  summarize(Percentile_75 = quantile(values,0.75),n_sup = sum(values > Percentile_75))

# # A tibble: 2 x 3
#       factor Percentile_75 n_sup
#        <chr>         <dbl> <int>
#   1 Factor.A           8.5     4
#   2 Factor.B           8.5     4

Upvotes: 2

Related Questions