From 22204f3bc9c9177763e76f517248be75d2d5a93b Mon Sep 17 00:00:00 2001
From: pan1fan2 <panfanubc@gmail.com>
Date: Sat, 13 Mar 2021 22:12:54 +0800
Subject: [PATCH] create eda r script

---
 R/eda.R                   | 34 +++++++++++++++++++--
 man/eda.Rd                |  8 ++++-
 tests/testthat/test-eda.R | 63 +++++++++++++++++++++++++++++++++++++++
 3 files changed, 102 insertions(+), 3 deletions(-)
 create mode 100644 tests/testthat/test-eda.R

diff --git a/R/eda.R b/R/eda.R
index baed54e..bd3d6c0 100644
--- a/R/eda.R
+++ b/R/eda.R
@@ -7,12 +7,42 @@
 #' @export
 #'
 #' @examples
-#' result <- eda(mtcars)
+#' df <- data.frame(num1 = c(8.5, 8, 9.2, 9.1, 9.4),
+#'                  num2 = c(0.88, 0.93, 0.95 , 0.92 , 0.91),
+#'                  num3 = c(0.46, 0.78, 0.66, 0.69, 0.52),
+#'                  num4 = c(0.082, 0.078, 0.082, 0.085, 0.066),
+#'                  cat1 = c("Good","Okay","Excellent","Terrible","Good"),
+#'                  target = c(2,2,3,1,3))
+#' result <- eda(df,"target")
 #' nb_num_feat  <- result$nb_num_features
 #' nb_cat_feat <- result$nb_cat_features
 #' ls_cat_feat <- result$cat_features
 #' ls_num_feat <- result$num_features
 #' nb_class<- result$nb_classes
 #' pair_plot <- result$pairplot
+
 eda <- function(df, target){
-}
\ No newline at end of file
+
+  results <- vector('list', 7)
+  names(results) <- c("num_features_name", "cat_features_name","nb_cat_features","nb_num_features","nb_class","missing","pairplot")
+
+  if (!is.data.frame(df)) {
+    stop('filter_wrapper expects a data frame object')
+  }
+  if(!(target %in% c(colnames(df)))) {
+    stop('column name is incorrect')
+
+  }
+
+  drop <- c(target)
+  df_fea <- df[,!(names(df) %in% drop)]
+  results$num_features_name = c(colnames(dplyr::select_if(df_fea, is.numeric)))
+  results$cat_features_name = c(colnames(df_fea[!(names(df_fea) %in% results$num_features_name)]))
+  results$nb_num_features = length(results$num_features_name)
+  results$nb_cat_features = length(results$cat_features_name)
+  results$nb_class = sort(unique(df[,target]))
+  results$missing = tibble::as_tibble(purrr::map(purrr::map_df(df_fea,is.na),sum))
+  results$pairplot = GGally::ggpairs(dplyr::select_if(df_fea, is.numeric))
+  return (results)
+
+}
diff --git a/man/eda.Rd b/man/eda.Rd
index fc61bd1..1765c94 100644
--- a/man/eda.Rd
+++ b/man/eda.Rd
@@ -18,7 +18,13 @@ eda(df, target)
 Run a quick EDA analysis
 }
 \examples{
-result <- eda(mtcars)
+df <- data.frame(num1 = c(8.5, 8, 9.2, 9.1, 9.4),
+                 num2 = c(0.88, 0.93, 0.95 , 0.92 , 0.91),
+                 num3 = c(0.46, 0.78, 0.66, 0.69, 0.52),
+                 num4 = c(0.082, 0.078, 0.082, 0.085, 0.066),
+                 cat1 = c("Good","Okay","Excellent","Terrible","Good"),
+                 target = c(2,2,3,1,3))
+result <- eda(df,"target")
 nb_num_feat  <- result$nb_num_features
 nb_cat_feat <- result$nb_cat_features
 ls_cat_feat <- result$cat_features
diff --git a/tests/testthat/test-eda.R b/tests/testthat/test-eda.R
new file mode 100644
index 0000000..822bc37
--- /dev/null
+++ b/tests/testthat/test-eda.R
@@ -0,0 +1,63 @@
+df <- data.frame(num1 = c(8.5, 8, 9.2, 9.1, 9.4),
+                       num2 = c(0.88, 0.93, 0.95 , 0.92 , 0.91),
+                       num3 = c(0.46, 0.78, 0.66, 0.69, 0.52),
+                       num4 = c(0.082, 0.078, 0.082, 0.085, 0.066),
+                       cat1 = c("Good","Okay","Excellent","Terrible","Good"),
+                       target = c(2,2,3,1,3))
+ls <- c(1,2,3,4,5)
+
+test_that('Plot should use gg matrix and the numerical column only', {
+  p <- eda(df,"target")$pairplot
+  expect_true("ggmatrix" %in% c(class(p)))
+  expect_true(eda(df,"target")$nb_num_features == p$ncol)
+})
+
+test_that('Check attributes', {
+  res <- eda(df,"target")
+  expect_true("pairplot" %in% names(res))
+  expect_true("nb_num_features" %in% names(res))
+})
+
+test_that('Test the length of numerical features', {
+  res <- eda(df,"target")
+  expect_true(res$nb_num_features == 4)
+})
+
+test_that('Test the length of categorical features', {
+  res <- eda(df,"target")
+  expect_true(res$nb_cat_features == 1)
+})
+
+test_that('Check numerical features names', {
+  res <- eda(df,"target")
+  expect_true("num1" %in% res$num_features_name)
+})
+
+test_that('Check categorical features names', {
+  res <- eda(df,"target")
+  expect_true("cat1" %in% res$cat_features_name)
+})
+
+test_that('Check class labels', {
+  res <- eda(df,"target")
+  expect_equivalent(res$nb_class, c(1,2,3))
+})
+
+test_that('Missing values',{
+  res <- eda(df,"target")
+  expect_equivalent(res$missing,
+                    data.frame(num1 = c(0),
+                              num2 = c(0),
+                              num3 = c(0),
+                              num4 = c(0),
+                              cat1 = c(0)))
+})
+
+# handle exception
+test_that('Check if user provides a data frame', {
+    expect_error(eda(ls,"target"))
+})
+
+test_that('Check if user provides a correct column name', {
+    expect_error(eda(df,"targ"))
+})