part3 #question 2

part3 #question 2.R

library(tidyverse)
library(readxl)

#load the training data
training_data <- read_excel("training_participation_survey.xlsx")

#question 1: Error rates in variables
cat("\nPART 3 - Question 1: Error Rates\n")

#check program status errors (should only be specific values)

valid_program_status <- c("Completed and Passed","Completed but Failed","Dropped Out")

program_status_errors <- sum(!training_data$program_status %in% valid_program_status) / nrow(training_data) * 100

cat("Program Status Error Rate: ", round(program_status_errors, 1),"%\n")

# scor de satisfactie 1-5
satisfaction_errors <- sum(training_data$training_hours < 1  | training_data$training_hours > 5, na_rm = TRUE) / nrow(training_data) * 100
cat("Satisfaction Score Error Rate: ", round(satisfaction_errors, 1),"%\n")

# 0 - 2000 interval rezonabil, tot ce este in afara acestui interval este considerat eroare
training_hours_errors <- sum(training_data$training_hours < 0  | training_data$training_hours > 2000, na_rm = TRUE) / nrow(training_data) * 100
cat("Training Hours Error Rate: ", round(training_hours_errors, 1),"%\n")


# vec <- c("Completed and Passed","Completed but Failed","Dropped Out")

#Question2

# clean the data for questions 2 - 4
traning_clean <- training_data %>%
 mutate(
      #fix program status spelling errors
      program_status_clean = case_when(
            str_detect(tolower(program_status), "completed.*passed") ~"Completed and Passed",
            str_detect(tolower(program_status), "completed.*failed") ~"Completed but Failed",
            str_detect(tolower(program_status), "dropped") - "Dropped Out",
            TRUE ~ program_status
      ),

      #fix satisfaction_score (1-5scale)
      satisfaction_score_clean = ifelse (
          satisfaction_score >= 1 & satisfaction_score <= 5,
          satisfaction_score,
          median(satisfaction_score[satisfaction_score >= 1 & satisfaction_score <= 5], na.rm = TRUE)
      ),

      #fix satisfaction training hours
      training_hours_clean = ifelse(
             training_hours >= 0 & training_hours <= 2000,
             training_hours,
             median(satisfaction_hours[training_hours >= 0 &  training_hours<= 2000], na.rm = TRUE)
      )
 )
 cat("PART 3 - Question 2: Mean Satisfaction by Program Status and Job Offer\n")

 satisfaction_analysis <- traning_clean %>%

 summarise(
     mean_satisfaction = mean(satisfaction_score_clean, na.rm = TRUE),
     .groups = "drop"
 ) %>%

 pivot_wider(

    names_from = job_offer_within_3mo,
    values_from = mean_satisfaction,
    names_prefix = "job_offer_"
 )

 print(satisfaction_analysis)