Programming Assignment 3

INTRODUCTION

LOAD HELPER FUNCTIONS

source('helper_functions.R')

Question 1

#View data set

outcome <- read.csv("data/outcome-of-care-measures.csv", colClasses = "character")

head(outcome)

dim(outcome)

[1] 4706   46

best <- function(state = 'tx', outcome = 'heart attack'){
  #Clean up white space
  state <- toupper(trimws(state)) 
  outcome <- tolower(trimws(outcome))
  
  #bestFx() is a helper function with various methods
  bestObj <- bestFx()
  
  #Read outcome data
  df <- bestObj$getData()
  
  #Check that state and outcome are valid
  states <- bestObj$states() #get all states
  outcomes <- bestObj$outcome() #get all possible outcomes:heart attack, heart...
  
  if(!state %in% states){
    stop("Invalid state!", call. = FALSE)
  }
  
  if(!outcome %in% outcomes){
    stop("Invalid outcome!", call. = FALSE)
    
  }
  
  #bestHosp: function to return hospital name in the state with lowest 30-day death rate
  bestObj$bestHosp(df, state, outcome) #
  
}

best('TX', 'heart failure')

[1] "FORT DUNCAN MEDICAL CENTER"

best('NY', 'hert attack')

Error: Invalid outcome!

best('md', 'pneumonia')

[1] "GREATER BALTIMORE MEDICAL CENTER"

best('kd', 'pneumonia')

Error: Invalid state!

  #####################################################################################
  # Function name: rankHospital()
  # - function to return hospital name in the state with the given rank
  # - Inputs: state, outcome, rank
  # - Output: -hospital name
  #####################################################################################
  rankHospital <- function(state1 = 'TX', outcome = 'heart attack', rank = c('best', 'worst')){
    
      state1 <- toupper(trimws(state1)) 
      outcome <- tolower(trimws(outcome))
  
      rank <- rank[1]
      
      #Convert rank to lower case, trim white space
      if(is.character(rank)){
        rank <- trimws(tolower(rank))
      }
      
      #Read in data
      bestObj <- bestFx() 
      df <- bestObj$getData()
    
      #Check validity of state
      states = bestObj$states()
      
      if(!state1 %in% states){
        stop("Invalid State!", call. = FALSE)
      }
      
      #Check validity of outcome
      outcomes <- bestObj$outcome()
      if(!outcome %in% outcomes){
        stop("Invalid Outcome!", call. = FALSE)
      }
      
      #Filter for appropriate state
      df <- df %>%
        filter(state == state1) %>%
        select(hospital, all_of(outcome), state) %>%  #https://tidyselect.r-lib.org/reference/faq-external-vector.html
        drop_na(all_of(outcome)) #Remove NAs from outcome column
      
      n <- dim(df)[1] #number of rows in dataset
      
      ranks <- seq(n) #ranks of each hospital in data set
      
      if(rank == 'best'){
        rank <- 1
      }
      if(rank == 'worst'){
        rank <- n
      } 
      if (rank > n){
        return(NA)
      }
      
      #sort by outcome then by hospital name
      df <- df[order(df[, 2], df[, 1]), ] %>%
        cbind(ranks) #add column to dataframe df
      
      df[rank, ]$hospital #get name of hospital
  
}

rankHospital(' tx', 'heart failure', 4)

[1] "DETAR HOSPITAL NAVARRO"

rankHospital('md', 'heart attack', 'worst')

[1] "HARFORD MEMORIAL HOSPITAL"

rankHospital('mn ', 'heart attack', 5000)

[1] NA

rankHospital('m n ', 'heart attack', 10)

Error: Invalid State!

rankHospital('mn ', 'heart attack', 10)

[1] "VIRGINIA REGIONAL MEDICAL CENTER"

rankall <- function(outcome = "heart attack", rank = 'best'){
  
  ## Read data
  bestObj <- bestFx()
  df <- bestObj$getData()
  
  ##Check that outcome is valid
  outcome <- trimws(tolower(outcome))
  outcomes <- bestObj$outcome()  #Get list of outcomes
  if(!outcome %in% outcomes){
    stop("Invalid input!")
  }
  
  ##For each state, find the hospital of the given rank
  states <- sort(bestObj$states())
  
  #Create empty dataframe
  d_f <- data.frame(hospital = character(),
                    state = character(),
                    stringsAsFactors = FALSE)
  
  for (state in states){
    
    #getHosp(): returns data frame with hospital and state as columns
    temp <- bestObj$getHosp(df, state, outcome, rank)

    d_f <- rbind(d_f, temp)
  }

  d_f
  
}

rankall('heart attack', 20)

tail(rankall('pneumonia' ,'worst'), 3)

tail(rankall('heart failure' ), 10)