Remove non-coders from proportion (#78)

best-practice-and-impact · Jan 11, 2024 · fd07cbf · fd07cbf
1 parent 9b9e6e1
commit fd07cbf
Show file tree

Hide file tree

Showing 2 changed files with 13 additions and 8 deletions.
diff --git a/R/frequency-tables.R b/R/frequency-tables.R
@@ -1144,6 +1144,9 @@ summarise_heard_of_RAP_by_prof <- function(data) {
 #' @export
 
 summarise_os_vs_prop <- function(data) {
+
+  data <- dplyr::filter(data, !is.na(code_freq), code_freq != "Never")
+
   data$open_source_lang_knowledge <- ifelse(
     data$knowledge_python == "Yes" | data$knowledge_R == "Yes",
     TRUE, FALSE

diff --git a/tests/testthat/test-summarise_os_vs_prop.R b/tests/testthat/test-summarise_os_vs_prop.R
@@ -1,10 +1,12 @@
 
-dummy_data <- data.frame(year = rep(c(2020, 2021, 2022), each=32),
-                         knowledge_python = rep(rep(c("Yes", "No"), each=16), times=3),
-                         knowledge_R = rep(rep(c("Yes", "No"), each=8), times=6),
-                         knowledge_SAS = rep(rep(c("Yes", "No"), each=4), times=12),
-                         knowledge_SPSS = rep(rep(c("Yes", "No"), each=2), times=24),
-                         knowledge_stata = rep(c("Yes", "No"), times=48))
+n=192
+dummy_data <- data.frame(year = rep(c(2020, 2021, 2022), each=n/3),
+                         code_freq = rep(c("Never", "Sometimes"), each=n/6, times=n/32),
+                         knowledge_python = rep(c("Yes", "No"), each=n/12, times=n/16),
+                         knowledge_R = rep(c("Yes", "No"), each=n/24, times=n/8),
+                         knowledge_SAS = rep(c("Yes", "No"), each=n/48, times=n/4),
+                         knowledge_SPSS = rep(c("Yes", "No"), each=n/96, times=n/2),
+                         knowledge_stata = rep(c("Yes", "No"), each=n/192, times=n))
 
 test_that("summarise_os_vs_prop missing data is handled correctly", {
 
@@ -23,8 +25,8 @@ test_that("summarise_os_vs_prop output is as expected", {
   expected <- data.frame(lang_type = factor(rep(c("Open Source", "Proprietary"), each=3),
                                             levels = c("Open Source", "Proprietary")),
                          year = rep(c("2020", "2021", "2022"), times=2),
-                         Freq = rep(c(24, 28), each=3),
-                         n = rep(32, times=6)) %>%
+                         Freq = rep(c(48, 56), each=3),
+                         n = rep(64, times=6)) %>%
     get_ci(freq_col = 3, n_col = 4)
 
   expect_equal(got, expected)