From c629848502edd4ce1da5ecba4b3433b99d050576 Mon Sep 17 00:00:00 2001
From: Kyle Belanger <kyleb44@hotmail.com>
Date: Mon, 9 Jan 2023 09:07:17 -0500
Subject: [PATCH] Create 1-data-exploration.R

---
 ML/1-data-exploration.R | 44 +++++++++++++++++++++++++++++++++++++++++
 1 file changed, 44 insertions(+)
 create mode 100644 ML/1-data-exploration.R

diff --git a/ML/1-data-exploration.R b/ML/1-data-exploration.R
new file mode 100644
index 0000000..8052b6d
--- /dev/null
+++ b/ML/1-data-exploration.R
@@ -0,0 +1,44 @@
+rm(list = ls(all.names = TRUE)) # Clear the memory of variables from previous run.
+cat("\014") # Clear the console
+
+
+# load packages -----------------------------------------------------------
+
+box::use(
+  magrittr[`%>%`]
+  ,here[here]
+  ,dplyr
+  ,readr
+  ,tidyr
+  ,ggplot2
+)
+
+
+
+
+# load data ---------------------------------------------------------------
+
+ds_high_tsh <-   readr$read_rds(
+    here("ML","data-unshared","ds_high_tsh.RDS")
+  )
+
+
+
+# data manipulation -------------------------------------------------------
+
+#here I am adding a column to determine if the Free T4 Value is diagnostic or not
+# using the FT4 Referance range low as the cut off (0.93)
+
+
+ds_high_tsh <- ds_high_tsh %>%
+  dplyr$mutate(ft4_dia = dplyr$if_else(`50995` < 0.93, 1, 0))
+
+
+
+# basic visualization -----------------------------------------------------
+
+g1 <- ds_high_tsh %>%
+  dplyr$select(-subject_id, - charttime) %>%
+  dplyr$mutate(dplyr$across(gender, ~dplyr$recode(.,M = 1, F = 2))) %>%
+  tidyr$pivot_longer(cols = dplyr$everything()) %>%
+