In BAREJAA/reactivity: This package provides a way to generate introductory dplyr and ggplot2 presentations using user-supplied tidy datasets

knitr::opts_chunk$set(echo = TRUE)
library(tidyverse)
library(here)
#library(babynames)

Import heart data

source(here::here("code", "healthcare.R")) #loads dataset-specific variables

Family history

# Create a "family_history" variable
fam_hist <- c("Aunt", "aunt", "Auntie", "auntie", "Father", "father", "Uncle", "uncle", "Grandmother", "grandmother", "Grandfather", "grandfather", "Mother", "mother")

heart$family_history <- sample(fam_hist, nrow(heart), replace = TRUE) # exercise idea: use Stringr to identify patients whose grandPARENTS have heart disease

# You could also show how to clean up names by standardising

Making a new dataframe

heart <- heart %>% 
          mutate(health_status = case_when(
            fbs == "elevated" & chol > 240 ~ "Diabetic & High Cholesterol",
            fbs == "elevated" ~ "Diabetic",
            chol > 240 ~ "High Cholesterol",
            TRUE ~ "Normal blood sugar and cholesterol"
          ))

Define drugs & fun()

#trestbps (high bp)
high_trestbps <- c("felodipine", "isradipine", "amlodipine", NA)

#chol (cholesterol)
high_chol <- c("atorvastatin", "simvastatin", "lovastatin", NA)

#fbs (diabetes)
high_fbs <- c("insulin", "metformin", "glyburide", "glimepiride", "rosiglitazone", NA)

#exang (angina)
high_exang <- c("aspirin", "nitrates", "Metoprolol", "Acebutolol", NA)

#target
#3 of the above

#functions
fill_trestbps <- function(var) {
  if (var > 120) {
    sample(high_trestbps, 1)
  } else {
    NA
  }
}
fill_chol <- function(var) {
  if (var > 240) {
    sample(high_chol, 1)
  } else {
    NA
  }
}
fill_fbs <- function(var) {
  if (var == "elevated") {
    sample(high_fbs, 1)
  } else {
    NA
  }
}
fill_exang <- function(var) {
  if (var == 1) {
    sample(high_exang, 1)
  } else {
    NA
  }
}

heart <- heart %>% 
   mutate(drug_trestpbs = map_chr(trestbps, fill_trestbps), 
          drug_chol = map_chr(chol, fill_chol), 
          drug_fbs = map_chr(fbs, fill_fbs), 
          drug_exang = map_chr(exang, fill_exang)) %>% 
   unite("medication_hx", starts_with("drug"), sep = ", ", remove = TRUE, na.rm = TRUE)

Quick checks

heart %>% count(health_status) 
heart %>% count(medication_hx, sort = TRUE)

make final df

phx <- heart %>% 
   select(patient_id, family_history, health_status, medication_hx)

write_csv(phx, here::here("data", "phx.csv"))

BAREJAA/reactivity documentation built on April 16, 2020, 6:57 p.m.