My current table looks like this:

This is the code I used to create that table.

CIregion_prop <- dta %>%
   filter(!is.na(Diabetes)) %>%
   filter(!is.na(region)) %>%
   group_by(region) %>%
   count(diabetes) %>%
   mutate(perc =prop.table(n)*100,
   lower = lapply(n, prop.test, n = sum(n)),
   upper = sapply(lower, function(x) x$conf.int[2])*100, 
   lower = sapply(lower, function(x) x$conf.int[1])*100)

I want to transform the table to below. Organized by how many people are positive and negative in each region to look like this:

How can I transform my above code?

CodePudding user response：

Try using dplyr::select() to remove the region data and omit the group_by() step:

library(dplyr)

region <- sample(c(1,2,3), 1649, replace = T)
Diabetes <- sample(c(0,1), 1649, replace = T)

df <- data.frame(region, Diabetes)

CI.no.region_prop <- df %>%
  filter(!is.na(Diabetes)) %>%
  filter(!is.na(region)) %>%
  dplyr::select(Diabetes) %>%
  #group_by(region) %>%
  count(Diabetes) %>%
  mutate(perc = prop.table(n)*100,
         lower = lapply(n, prop.test, n = sum(n)),
         upper = sapply(lower, function(x) x$conf.int[2])*100, 
         lower = sapply(lower, function(x) x$conf.int[1])*100)