TUS 2024 I.Dönem Sonuçları (Kurum Bazında)

R Codes

#| label: Import libraries

# Define the required package names
packages <- c("tidyverse", 
              "here", 
              "pdftools", 
              "glue", 
              "ggtext")

# Function to check, install, and load packages
load_packages <- function(packages) {
        
  for (package in packages) {
    if (!require(package, character.only = TRUE)) {
        install.packages(package)
        library(package, character.only = TRUE)
    } else {
      suppressPackageStartupMessages(library(package, character.only = TRUE))
        # library(package, character.only = TRUE)
    }
  }
}

load_packages(packages)

kadro_tus <- pdftools::pdf_text(pdf = here::here("datasets", "minmax_td03062024.pdf")) %>%
  as_tibble()

kadro_tus_init <- kadro_tus %>%
  separate_rows(value, sep = "\n") %>%
  mutate(value = str_squish(value)) %>%
  filter(value != "") %>%
  mutate(value = str_remove_all(value, " dipnot \\d")) %>%
  filter(!str_detect(value, "Kontenjan|2024-TUS")) %>%
  extract(value, into = c("kurum", "b"), "(.*)/(.*)") %>%
  extract(b, into = c("brans", "tur", "d"), "(.*)(Yabancı Uyruklu|Genel)(.*)") %>%
  mutate(
    d = str_trim(d),
    brans = str_trim(brans)
  ) %>%
  separate(d, into = c("kontenjan", "yerlesen", "bos", "min", "max"), sep = " ") %>%
  separate(kurum, into = c("number", "kurum"), sep = "^\\d+\\s") %>%
  select(-number) %>%
  mutate(
    kurum = str_replace_all(kurum, "Sağlık Bilimleri Üniversitesi", "SBÜ"),
    kurum = str_replace_all(kurum, "Eğitim ve Araştırma Hastanesi", "EAH"),
    kurum = str_replace_all(kurum, "Tıp Fakültesi", "Tıp F"),
    kurum = str_remove_all(kurum, " \\(ANKARA\\)"),
    kurum = str_remove_all(kurum, " \\(İSTANBUL\\)"),
    kurum = str_remove_all(kurum, "Tayfur Ata Sökmen "),
    kurum = str_remove_all(kurum, "Prof. Dr. ")
  ) %>%
  mutate(across(min:max, ~ if_else(. == "--", NA_character_, .))) %>%
  mutate(across(min:max, ~ str_replace_all(., ",", "."))) %>%
  mutate(across(kontenjan:max, ~ as.numeric(.))) %>%
  mutate(across(min:max, ~ round(., 2))) %>%
  mutate(
    min = if_else(is.na(min), 0, min),
    doluluk = 100 * round(yerlesen / kontenjan, 2),
    kurum_title = if_else(!str_detect(kurum, "SBÜ|EAH"), str_to_title(kurum), kurum),
    equal_kucuk_puan = if_else(min != max, min, NA_real_)
  ) %>%
  filter(
    brans != "ASKERİ SAĞLIK HİZMETLERİ",
    brans != "HAVA VE UZAY HEKİMLİĞİ",
    tur == "Genel"
  ) %>%
  mutate(
    brans_title = str_to_title(brans),
    brans_title = str_replace_all(brans_title, "i̇", "i"),
    brans_title = str_replace_all(brans_title, "Hastaliklari", "Hastalıkları"),
    brans_title = str_replace_all(brans_title, "Kadin", "Kadın"),
    brans_title = str_replace_all(brans_title, "Sağliği", "Sağlığı"),
    brans_title = str_replace_all(brans_title, "Ve ", "ve "),
    brans_title = str_replace_all(brans_title, " Tip", " Tıp"),
    brans_title = str_replace_all(brans_title, "Tibbi ", "Tıbbi ")
  ) %>%
  filter(!str_detect(kurum, "Dr. Sami Ulus")) %>%               # Empty value, causes error in function
  filter(!str_detect(kurum, "Jandarma Genel Komutanlığı"))  %>%    # Empty value, causes error in function 
  filter(!str_detect(kurum, "Giresun Kadın Doğum ve Çocuk")) # Empty value, causes error in function

my_grays <- c("#C0C0C0", "#909090", "#696969", "#525252")
grid_color <- c("#E8E8E8")
empty_color <- c("#A12123")
some_empty_color <- c("#919498")
ktu_color <- c("#003B64", "#1B71AC") # I only know the KTU's official colors. If any database exist, would be great.


# Function to generate a random color in hex format
generate_random_hex_color <- function(seed) {
  # Set seed for reproducibility
  set.seed(seed)

  # Generate random values for red, green, and blue components
  red <- sprintf("%02X", sample(0:255, 1))
  green <- sprintf("%02X", sample(0:255, 1))
  blue <- sprintf("%02X", sample(0:255, 1))

  # Combine components into a hex color code
  hex_color <- paste0("#", red, green, blue)

  return(hex_color)
}


# Function to generate dark and light versions of a color
generate_dark_light_colors <- function(base_color, dark_factor = 0.7, light_factor = 1.3) {
  # Convert the hexadecimal color to RGB values
  base_rgb <- col2rgb(base_color)

  # Calculate dark and light variations
  dark_rgb <- round(base_rgb * dark_factor)
  light_rgb <- round(base_rgb * light_factor)

  # Ensure values are within valid RGB range (0-255)
  dark_rgb <- pmax(0, pmin(255, dark_rgb))
  light_rgb <- pmax(0, pmin(255, light_rgb))

  # Convert RGB values back to hexadecimal
  dark_color <- rgb(dark_rgb[1], dark_rgb[2], dark_rgb[3], maxColorValue = 255)
  light_color <- rgb(light_rgb[1], light_rgb[2], light_rgb[3], maxColorValue = 255)

  return(list(dark_color = dark_color, light_color = light_color))
}


## Create color list

unique_kurums <- unique(kadro_tus_init$kurum)           # List of all institutes

num_colors <- length(unique_kurums)                     # required number of base colors
base_colors <- replicate(num_colors, generate_random_hex_color(seed = NULL), simplify = FALSE)

my_color_pallette <- map(base_colors, ~ {
  generated_colors <- generate_dark_light_colors(.x)

  return(list(
    random_base_color = .x,
    dark_color = generated_colors$dark_color,
    light_color = generated_colors$light_color
  ))
})


## Create institute df with color list

color_for_kurums <- list()

for (i in seq_along(my_color_pallette)) {
  kurum <- as_tibble(unique_kurums)[[1]][[i]]
  random_base_color <- my_color_pallette[[i]]$random_base_color
  dark_color <- my_color_pallette[[i]]$dark_color
  light_color <- my_color_pallette[[i]]$light_color

  single_kurum_df <- data.frame(kurum, random_base_color, dark_color, light_color)

  # Append the current data frame to the result list
  color_for_kurums[[i]] <- single_kurum_df
}


## Bind list and Change only KTU colors
color_for_kurums_df <- list_rbind(color_for_kurums) %>%
  mutate(
    random_base_color = if_else(kurum == "Karadeniz Teknik Üniversitesi Tıp F", ktu_color[1], random_base_color),
    dark_color = if_else(kurum == "Karadeniz Teknik Üniversitesi Tıp F", ktu_color[1], dark_color),
    light_color = if_else(kurum == "Karadeniz Teknik Üniversitesi Tıp F", ktu_color[2], light_color)
  )

Divide dataset into three part for Üniversite - EAH - Sehir Hastanesi

## To remove some errors in ÖSYM dataset
kadro_tus_init_colored <- kadro_tus_init %>% 
    group_by(brans, kurum) %>% 
    arrange(desc(kontenjan)) %>%
    slice(1) %>%
    ungroup() %>% 
  left_join(color_for_kurums_df, by = "kurum")

## Define sub datasets for Top, EAH, Şehir Hastanesi

data_uni <- kadro_tus_init_colored %>% 
  filter(str_detect(kurum, "Tıp F"))

data_eah <- kadro_tus_init_colored %>% 
  filter(str_detect(kurum, " EAH"))

data_sehir <- kadro_tus_init_colored %>% 
  filter(str_detect(kurum, "Şehir Hastanesi"))

Custom FUNCTION to plot (Main plot)

func_tus_institute <- function(selected_dataset, my_institute) {
  
  clean_institute_2024 <- selected_dataset %>%
    filter(
      kurum == {{ my_institute }}
    ) %>%
    group_by(brans) %>%
    arrange(desc(kontenjan)) %>%
    slice(1) %>%
    ungroup()

  minmax_branch <- clean_institute_2024 %>%
    filter(min != 0) %>%
    summarise(
      min = min(min, na.rm = TRUE),
      max = max(max, na.rm = TRUE)
    )

  doluluk_branch <- clean_institute_2024 %>%
    summarise(
      sum_kontenjan = sum(kontenjan, na.rm = TRUE),
      sum_yerlesen = sum(yerlesen, na.rm = TRUE)
    ) %>%
    mutate(doluluk = paste0("%", round(100 * sum_yerlesen / sum_kontenjan, 1)))


  lowest <- clean_institute_2024 %>%
    filter(min != 0) %>%
    arrange(min) %>%
    slice_head() %>%
    pull(brans_title)

  highest <- clean_institute_2024 %>%
    filter(min != 0) %>%
    arrange(desc(max)) %>%
    slice_head() %>%
    pull(brans_title)


  
  my_light_color <- clean_institute_2024 %>%
    slice_head() %>%
    pull(light_color)
  my_dark_color <- clean_institute_2024 %>%
    slice_head() %>%
    pull(dark_color)


  my_title <- glue:::glue("TUS 2024/I - <span style='color:{my_dark_color};'>{my_institute}</span>")
  my_subtitle <- glue::glue("Doluluk: {doluluk_branch$doluluk} ({doluluk_branch$sum_yerlesen}/{doluluk_branch$sum_kontenjan})<br><br> 
                            En yüksek: **{highest}** ({minmax_branch$max})<br>")
  my_caption <- glue::glue("<i style='color:{my_dark_color}'>Ana Renk</i>: Kadrolar tam dolu<br>kaynak: ÖSYM<br>@AliGunerMD")


  institute_plot <- clean_institute_2024 %>%
    # ggplot(aes(kontenjan, yerlesen)) +
    # geom_point()
    mutate(
      color = case_when(
        doluluk == 100 ~ my_dark_color,
        doluluk == 0 ~ empty_color,
        TRUE ~ some_empty_color
      ),
      name = glue("<i style='color:{color}'>{brans_title}</i> ({yerlesen}\\/{kontenjan})"),
      mod_min = if_else(yerlesen == 0, -kontenjan, min + 100),
      name = fct_reorder(name, mod_min)
    ) %>%
    ggplot(aes(x = min, y = name, fill = color)) +
    geom_segment(aes(x = min, xend = max, y = name, yend = name), color = my_grays[3]) +
    geom_point(aes(x = min, y = name), color = ifelse(clean_institute_2024$doluluk == 100, my_light_color, my_grays[1]), size = 2.5) +
    geom_point(aes(x = max, y = name), color = ifelse(clean_institute_2024$doluluk == 100, my_dark_color, my_grays[2]), size = 3.5) +
    theme_light() +
    scale_x_continuous(limits = c(minmax_branch$min - 2, minmax_branch$max + 2), breaks = seq(40, 80, 5)) +
    geom_text(aes(x = max, y = name, label = max),
      hjust = -.5, size = 3,
      color = ifelse(clean_institute_2024$doluluk == 100, my_dark_color, my_grays[2])
    ) +
    geom_text(aes(x = equal_kucuk_puan, y = name, label = equal_kucuk_puan), hjust = 1.5, size = 2.5, color = ifelse(clean_institute_2024$doluluk == 100, my_dark_color, my_grays[2]), alpha = .7, check_overlap = TRUE) +
    labs(
      x = "Puan",
      y = "",
      title = my_title,
      caption = my_caption,
      subtitle = my_subtitle
    ) +
    theme(
      panel.grid = element_line(color = grid_color),
      panel.grid.minor.x = element_blank(),
      # panel.grid.major.y = element_line(color = my_light_color, size = .1),
      panel.border = element_blank(),
      axis.text.y = element_markdown(face = "bold"),
      legend.position = "none",
      plot.caption = element_markdown(),
      plot.title.position = "plot",
      plot.title = element_markdown(hjust = .5, size = 14, face = "bold"),
      plot.subtitle = element_markdown(hjust = .5)
    )

  # print(clean_institute_2024$kurum_title)       # used for debugging
  # print(highest)                                # used for debugging
  # print(lowest)                                 # used for debugging


  return(institute_plot)
}

Custom FUNCTION to use func_tus_institute() function for all DATASETS and all Institutes (and to save)

datasets <- list(
  data_sehir = data_sehir,
  data_eah = data_eah,
  data_uni = data_uni
)

# Create the main output folder
# output_folder <- "content/post/2024-06-04-tus-24-1-uni/Outputs_2024_1"    # to prepare
output_folder <- "Outputs_2024_1"                                           # to render

dir.create(output_folder, showWarnings = FALSE)


# Iterate over datasets
for (dataset_name in names(datasets)) {
  dataset <- datasets[[dataset_name]]

  # Create a folder for each dataset within the main output folder
  dataset_folder <- file.path(output_folder, dataset_name)
  dir.create(dataset_folder, showWarnings = FALSE)

  # Get unique values of 'kurum' in the current dataset
  institute_names <- distinct(dataset, kurum) %>%
    pull(kurum)

  # Iterate over unique 'kurum' values
  for (institute in institute_names) {
    # Call the function for the current dataset and 'kurum' value
    institute_plot <- func_tus_institute(selected_dataset = dataset, my_institute = institute)

    # Save the plot in the 'kurum' folder
    plot_filename <- file.path(paste0(dataset_folder, "/", as.character(institute), ".jpg"))

    # to select save height dynamically
    no_brans <- dataset %>%
      filter(kurum == {{ institute }}) %>%
      distinct(brans) %>%
      nrow()

    my_height <- case_when(
      no_brans >= 30 ~ 12,
      no_brans >= 20 & no_brans < 30 ~ 11,
      no_brans >= 10 & no_brans < 20 ~ 10,
      no_brans >= 5 & no_brans < 10 ~ 6,
      no_brans < 5 ~ 4,
      TRUE ~ 11
    )

    # Save plot
    ggsave(
      file = plot_filename,
      plot = institute_plot,
      dpi = 300,
      width = 9,
      height = 11
    )
  }
}