5  Hubungan Antar Peubah

5.1 Data

Data yang akan digunakan adalah

Code 
# Add Google Fonts 
font_add_google("Lato", "lato")
showtext_auto()

# Data
data_cars <- mtcars %>%
  mutate(
    cyl = as.factor(cyl),   
    gear = as.factor(gear), 
    carb = as.factor(carb),
    cc = disp * 16.3871, # konversi volume mesin ke cc
    engine = factor(ifelse(vs == 0, "V-Engine", "Inline")),
    trans = factor(ifelse(am == 0, "Automatic", "Manual"))
  ) %>% 
  rownames_to_column(var = "car_name") %>% 
  select(-disp, -vs, -am) 
  
datatable(data_cars)

6 Relationship

Hubungan antar variabel dalam data dapat divisualisasikan menggunakan berbagai teknik yang membantu memahami korelasi dan pola tersembunyi.

6.1 Scatterplot

Scatterplot digunakan untuk menunjukkan hubungan antara dua variabel numerik.

Code 
# Pilih 3 mobil yang akan di-highlight
label_cars <- bind_rows(
  # Mobil paling irit
  data_cars %>% slice_max(mpg, n = 1, with_ties = FALSE), 
  
  # Mobil menengah
  data_cars %>% 
    mutate(dist_to_median = abs(mpg - median(mpg))) %>%   
    slice_min(dist_to_median, n = 1, with_ties = FALSE) %>%  
    select(-dist_to_median),  
  
  # Mobil paling boros
  data_cars %>% slice_min(mpg, n = 1, with_ties = FALSE)  
) %>%
  arrange(desc(mpg)) 

colors <- c("#21637b", "#967349", "#ad0c1b")

6.1.1 Advanced

Code 
# viz
chart <-
ggplot(data_cars,
       aes(x = cc, y = mpg, 
           fill = factor(cyl), color = factor(cyl), 
           shape = factor(trans), size = hp)) +  
  
  # Scatterplot
  geom_point(stroke = 1, alpha = 0.5) +  
  
  # Aes
  scale_shape_manual(values = c(21, 23)) +
  scale_size_continuous(range = c(1, 10)) +
  scale_fill_manual(values = colors) +  
  scale_color_manual(values = colors) +
  
  # Label
  geom_text_repel(
    data = label_cars, aes(label = car_name), 
    size = 6, fontface = "bold", color = "black", 
    box.padding = 0.5, point.padding = 0.5, 
    nudge_x = 15, 
    nudge_y = 2,
    segment.color = "black", segment.size = 0.5, 
    arrow = arrow(length = unit(0.01, "npc"))
  ) +
  
  # Setting 
  labs(title = "Efek Mesin & Transmisi terhadap Efisiensi BBM",
       subtitle = "(Ukuran simbol menunjukkan tenaga mesin (hp))",
       x = "Kapasitas Mesin (cc)",
       y = "← Boros          Efisiensi BBM          Irit →",
       color = "Jumlah Silinder") +
  
  guides(
    size = "none", 
    fill = "none",
    shape = guide_legend(title = NULL)
  ) + 
  
  theme_minimal(base_size = 35) + # Global Font size
  theme(
    axis.title.y = element_text(size = 30, face = "bold"),
    axis.text.y = element_text(hjust = 1, face = "bold", 
                               margin = margin(r = 1)),
    plot.title = element_text(hjust = -0.2, face = "bold",
                              margin = margin(b = 5)),
    plot.subtitle = element_text(hjust = -0.115, size=25),
    
    # Garis
    panel.grid.major.y = element_blank(),
    panel.grid.minor.y = element_blank(),
    panel.grid.major.x = element_blank(),
    panel.grid.minor.x = element_blank(),
    axis.line = element_line(linewidth = 0.5, colour = "black"),
    
    # Legend
    legend.title = element_text(size = 30),
    legend.position = "top",
    legend.direction = "horizontal",
    legend.justification = c(-0.25, 0),
    legend.margin = margin(t = -10, b = -25), 
    legend.text = element_text(margin = margin(r = -1, l = -1))
  ) 

chart

Code 
# Simpan Chart
ggsave("Chart/01_Scatterplot.png", chart, dpi = 300, height = 5, width = 9)

6.1.2 Sederhana

Code 
ggplot(data_cars, aes(x = cc, y = mpg)) +  
  # Scatterplot
  geom_point() +
  theme_gray(base_size = 35) #efek pake font google

6.2 Parallel Coordinates Plot

Visualisasi ini digunakan untuk menganalisis hubungan beberapa variabel numerik sekaligus.

Code 
# Pilih kolom
selected_columns <- c("mpg", "hp", "drat", "wt", "qsec", "cc")

new_labels <- c("Efisiensi BBM", "Tenaga", "Rasio Driveshaft", 
                "Berat Mobil", "Waktu 1/4 Mil", "Kapasitas Mesin")

# HIGHTLIGHT---------------------------------
highlight_cars <- function(data, col) {
  selected <- bind_rows(
    # Mobil dengan nilai tertinggi
    data %>% slice_max({{col}}, n = 1, with_ties = FALSE),  
    
    # Mobil dengan nilai median terdekat
    data %>% mutate(dist_to_median = abs({{col}} - median({{col}}))) %>% 
      slice_min(dist_to_median, n = 1, with_ties = FALSE) %>% 
      select(-dist_to_median),  
    
    # Mobil dengan nilai terendah
    data %>% slice_min({{col}}, n = 1, with_ties = FALSE)  
  ) %>%
    arrange(desc({{col}}))  
  
  return(selected)
}

# Pilih 3 mobil yang akan di-highlight
label_cars <- highlight_cars(data_cars, mpg)  # Berdasarkan efisiensi BBM

# Tambahkan kolom highlight dan alpha transparansi
data_cars <- data_cars %>%
  mutate(
    highlight = ifelse(car_name %in% label_cars$car_name, car_name, "Lainnya"),
    highlight = factor(highlight, levels = c(label_cars$car_name, "Lainnya")),
    alpha_val = ifelse(highlight == "Lainnya", 0.2, 1)
  )

# Warna khusus: 3 mobil utama + lainnya dalam warna abu-abu
highlight_colors <- c("#21637b", "#967349", "#ad0c1b", "#E8E8E8")  
color_mapping <- setNames(highlight_colors, levels(data_cars$highlight))


# NORMALISASI DATA ---------------------------------
# **Normalisasi data ke skala 0-1 secara manual**
data_cars_norm <- data_cars %>%
  mutate(across(all_of(selected_columns), rescale))

# **Pastikan label_positions memiliki skala yang sudah dinormalisasi**
label_positions <- data_cars_norm %>%
  dplyr::filter(car_name %in% label_cars$car_name) %>%
  select(car_name, cc, highlight)  # Tambahkan highlight agar sesuai dengan warna

x_max <- length(selected_columns) + 0.3  # Posisi label di ujung sumbu

6.2.1 Advanced 1

Highlight 3 mobil berdasarkan Efisiensi bahan bakar.

Code 
# PLOT
chart <- 
ggparcoord(data_cars_norm, 
           columns = match(selected_columns, colnames(data_cars_norm)),   
           groupColumn = "highlight",          
           scale = "uniminmax",  
           showPoints = TRUE,    
           alphaLines = 0.6) +   
  
  # Warna
  scale_color_manual(values = color_mapping, guide = "none") +  
  scale_alpha_identity() +  
  
  # TEXT ----------------
  # Judul
  labs(title = "Perbandingan Karakteristik Mobil Berdasarkan Efisiensi BBM",
       x = NULL,  
       y = NULL) +
  
  # Ganti label sumbu x
  scale_x_discrete(labels = new_labels) +

  # **Tambahkan label di ujung garis & geser lebih ke kiri**
  geom_text(data = label_positions, 
            aes(x = x_max - 0.2, y = cc, label = car_name, color = highlight), 
            fontface = "bold", size = 8, hjust = 0, inherit.aes = FALSE) +

  
  # THEME -------------------
  # Hapus Legend
  guides(color = "none") +

  # **Styling tema**
  theme_minimal(base_size = 35) +  
  theme(
    axis.title.y = element_text(size = 24, face = "bold"),
    axis.text.y = element_text(hjust = 1, face = "bold", 
                               margin = margin(r = -30), size = 22),
    plot.title = element_text(hjust = 0, face = "bold", size = 30),

    # **Pertebal grid utama**
    panel.grid.major = element_line(linewidth = 0.5),  
    panel.grid.minor = element_line(linewidth = 0.3),
    panel.grid.major.y = element_blank(),
    panel.grid.minor.y = element_blank()
  ) + 
  coord_cartesian(clip = "off")

chart

Code 
# Simpan Chart
ggsave("Chart/02_Paralell-Coordinates.png", 
       chart, dpi = 300, height = 5, width = 9)

6.2.2 Advanced 2

Hightlight mobil berdasarkan tenaga.

Code 
# Pilih 3 mobil yang akan di-highlight
label_cars <- highlight_cars(data_cars, hp)  # Berdasarkan tenaga

# Tambahkan kolom highlight dan alpha transparansi
data_cars <- data_cars %>%
  mutate(
    highlight = ifelse(car_name %in% label_cars$car_name, car_name, "Lainnya"),
    highlight = factor(highlight, levels = c(label_cars$car_name, "Lainnya")),
    alpha_val = ifelse(highlight == "Lainnya", 0.2, 1)
  )

# Warna khusus: 3 mobil utama + lainnya dalam warna abu-abu
highlight_colors <- c("#21637b", "#967349", "#ad0c1b", "#E8E8E8")  
color_mapping <- setNames(highlight_colors, levels(data_cars$highlight))


# NORMALISASI DATA ---------------------------------
# **Normalisasi data ke skala 0-1 secara manual**
data_cars_norm <- data_cars %>%
  mutate(across(all_of(selected_columns), rescale))

# **Pastikan label_positions memiliki skala yang sudah dinormalisasi**
label_positions <- data_cars_norm %>%
  dplyr::filter(car_name %in% label_cars$car_name) %>%
  select(car_name, cc, highlight)  # Tambahkan highlight agar sesuai dengan warna

x_max <- length(selected_columns) + 0.3  # Posisi label di ujung sumbu
Code 
# PLOT
chart <- 
ggparcoord(data_cars_norm, 
           columns = match(selected_columns, colnames(data_cars_norm)),   
           groupColumn = "highlight",          
           scale = "uniminmax",  
           showPoints = TRUE,    
           alphaLines = 0.6) +   
  
  # Warna
  scale_color_manual(values = color_mapping, guide = "none") +  
  scale_alpha_identity() +  
  
  # TEXT ----------------
  # Judul
  labs(title = "Perbandingan Karakteristik Mobil Berdasarkan Tenaga",
       x = NULL,  
       y = NULL) +
  
  # Ganti label sumbu x
  scale_x_discrete(labels = new_labels) +

  # **Tambahkan label di ujung garis & geser lebih ke kiri**
  geom_text(data = label_positions, 
            aes(x = x_max - 0.2, y = cc, label = car_name, color = highlight), 
            fontface = "bold", size = 8, hjust = 0, inherit.aes = FALSE) +

  
  # THEME -------------------
  # Hapus Legend
  guides(color = "none") +

  # **Styling tema**
  theme_minimal(base_size = 35) +  
  theme(
    axis.title.y = element_text(size = 24, face = "bold"),
    axis.text.y = element_text(hjust = 1, face = "bold", 
                               margin = margin(r = -30), size = 22),
    plot.title = element_text(hjust = 0, face = "bold", size = 30),

    # **Pertebal grid utama**
    panel.grid.major = element_line(linewidth = 0.5),  
    panel.grid.minor = element_line(linewidth = 0.3),
    panel.grid.major.y = element_blank(),
    panel.grid.minor.y = element_blank()
  ) + 
  coord_cartesian(clip = "off")

chart

Code 
# Simpan chart
ggsave("Chart/02_Paralell-Coordinates-2.png", 
       chart, dpi = 300, height = 5, width = 9)

6.2.3 Sederhana

Code 
ggparcoord(data_cars, 
           columns = match(selected_columns, colnames(data_cars)),  
           groupColumn = "cyl",
           scale = "uniminmax") +
  theme_gray(base_size = 35)

6.3 Radar Charts

Sering digunakan untuk membandingkan beberapa kategori dengan beberapa atribut dalam satu grafik.

6.3.1 Sederhana

Code 
ggRadar(data=iris,aes(group=Species))

Code 
ggRadar(data=mtcars,interactive=TRUE)
Code 
ggRadar(data=mtcars,aes(colour=am,facet=cyl),interactive=TRUE)
Code 
ggRadar(iris, aes(x=c(Sepal.Length, Sepal.Width, Petal.Length, Petal.Width)))

6.4 Chord Diagram

Diagram yang digunakan untuk menunjukkan hubungan antara beberapa kategori yang saling berhubungan.

Install Package Manual

  • Buka link ini di browser:
    👉 https://cran.r-project.org/bin/windows/contrib/4.2/

  • Cari circlize_0.4.16.zip atau versi terbaru.

  • Download file tersebut ke komputer Anda.

  • install.packages("C:/Users/NAMA_USER/Downloads/circlize_0.4.16.zip", repos = NULL, type = "win.binary")

Code 
# Install dan load library jika belum ada
# install.packages("circlize")
# Kalo gak bisa, coba ini
# install.packages("C:/Users/NAMA_USER/Downloads/circlize_0.4.16.zip", repos = NULL, type = "win.binary")

library(circlize)

# Load dataset
data("mtcars")

# Membuat data hubungan antara jumlah silinder dan jenis transmisi
mtcars$group <- as.character(mtcars$cyl)  # Mengubah cyl menjadi faktor karakter
mtcars$transmission <- ifelse(mtcars$am == 0, "Automatic", "Manual")

# Hitung frekuensi kombinasi
chord_data <- as.data.frame(table(mtcars$group, mtcars$transmission))

# Ubah nama kolom agar lebih mudah dipahami
colnames(chord_data) <- c("Cylinders", "Transmission", "Count")

# Buat chord diagram
chordDiagram(chord_data, transparency = 0.5, col = c("red", "blue", "green"))

6.5 Arc Chart

Menampilkan koneksi antar elemen dalam dataset dengan menggunakan kurva.

6.6 Correlation Matrix

Digunakan untuk menunjukkan hubungan antara banyak variabel dalam bentuk matriks.

6.7 Network Diagrams

Visualisasi hubungan antar entitas dalam bentuk jaringan.

6.8 Tree Diagrams

Menampilkan hubungan hierarkis dalam data.