Creating t-test confidence intervals using the new group_map

You need the in dev version of dplyr

#library(devtools)
#install_github("tidyverse/dplyr")

library(tidyverse)
library(broom)

calculate_ci <- function(d, x, ...) {
  d %>%
    group_by(...) %>%
    group_map(~t.test(.x %>% pull(!!enquo(x)), conf.int = TRUE) %>% tidy()) 
}

In the data mtcars, we group mpg by gear and am and calculate the confidence intervals

mtcars %>% calculate_ci(mpg, gear, am)