CODEWITHSUNDEEP
rdplyrpurrr
bfoste01
bfoste01

Reputation: 359

Generating a values for functions created with purrr:map

I'm trying to use purrr:map to create empirical cumulative percentages for values of x_var in a df that are unique to each level of a factor variable.

Ideally, I'd like a result to be a long df where the columns look as follows:

levels (long) | x_var | epcd_val

Here's an example:

# load packs
if(!require("pacman"))install.packages("pacman")
p_load(dplyr, tibble, purrr)

# generate fake data
samp_dat <- tibble(
x_var = rnorm (1000, 0, 1),
levels = sample(LETTERS[1:4], 1000, replace=TRUE, prob=c(0.25, 0.50, 0.125, 0.125)))

# generates a list of ecdf functions for each level
ecdfs <- samp_dat %>%
  group_split(levels) %>%
  map(., ~ ecdf(.x$x_var))

The resulting ecdfs is a list of edcf functions, which is unique to each level within levels.

I somehow then need to feed the x_var values, grouped on levels, back into this function. However, I'm stuck on how to pull it off with pipes.

Upvotes: 2

Views: 113

Answers (1)

akrun
akrun

Reputation: 887118

The ecdf outputs a function, so we feed the 'x_var' into the output function after grouping by 'levels'

library(dplyr)
samp_dat %>% 
    group_by(levels) %>%
    mutate(newval = ecdf(x_var)(x_var))

Upvotes: 3

Related Questions