# Summary statistik numerik
numeric_summary <- customer_data %>%
  select(Age, TotalPurchase, NumberOfVisits, FeedbackScore) %>%
  psych::describe() %>%
  select(n, mean, sd, min, max, median)

# Fungsi Statistik Utama 
get_stats <- function(x) {
  # hapus nilai NA
  x_no_na <- x[!is.na(x)]
  
  # mode_val akan menjadi NA jika vektor kosong atau mfv gagal
  mode_val <- tryCatch({
    # Gunakan mfv() dari package modeest untuk modus
    as.numeric(modeest::mfv(x_no_na)[1])
    }, error = function(e) {
      NA 
    })

  # Pastikan mean, median, dan SD dihitung dengan aman
  if (length(x_no_na) == 0) {
      return(data.frame(Mean = NA, Median = NA, Mode = NA, SD = NA))
  }

  # Output selalu 1 baris
  return(data.frame(
    Mean    = mean(x_no_na),
    Median = median(x_no_na),
    Mode    = mode_val,
    SD      = sd(x_no_na)
  ))
}

# Tampilkan summary
kable(numeric_summary, caption = "Statistik Deskriptif Variabel Numerik") %>%
  kable_styling(bootstrap_options = c("striped", "hover"), full_width = FALSE)

# Summary variabel kategorikal
cat_summary <- customer_data %>%
  select(Gender, StoreLocation, ProductCategory) %>%
  tidyr::gather(key = "Variable", value = "Category") %>%
  group_by(Variable, Category) %>%
  summarise(Count = n(), .groups = "drop") %>%
  mutate(Percentage = paste0(round(Count/nrow(customer_data)*100, 1), "%"))

kable(cat_summary, caption = "Distribusi Variabel Kategorikal") %>%
  kable_styling(bootstrap_options = c("striped", "hover"))

age_stats <- customer_data %>%
  summarise(
    Mean = round(mean(Age), 1),
    Median = median(Age),
    # PERUBAHAN: Menggunakan modeest::mfv()
    Mode = as.numeric(modeest::mfv(Age)[1]),
    SD = round(sd(Age), 1),
    Min = min(Age),
    Max = max(Age)
  )

# Hitung jarak antara mean dan median
gap <- abs(age_stats$Mean - age_stats$Median)

# Visualisasi distribusi umur 
p1 <- ggplot(customer_data, aes(x = Age)) +
  geom_histogram(aes(y = after_stat(density)), bins = 15, fill = "#349", alpha = 0.7) +
  geom_density(color = "#2c3e50", linewidth = 1) +
  geom_vline(xintercept = age_stats$Mean, color = "#e74c3c", linewidth = 1, linetype = "dashed") +
  geom_vline(xintercept = age_stats$Median, color = "#2ecc71", linewidth = 1, linetype = "dashed") +
  annotate("label",
           x = age_stats$Mean,
           y = 0.03,                     
           label = "Mean",
           color = "#e74c3c",
           fill = "white",
           alpha = 0.6,
           angle = 90,
           vjust = 1.2,
           hjust = 1.2,
           size = 4.5,
          label.size = 0) +
  annotate("label",
           x = age_stats$Median,
           y = 0.01,                    
           label = "Median",
           color = "#2ecc71",
           fill = "white",
           alpha = 0.6,
           angle = 90,
           vjust = -0.5,
           hjust = 1.2,
           size = 4.5,
          label.size = 0) +
  labs(title = "Distribusi Umur Pelanggan",
       x = "Umur", y = "Densitas") +
  theme_minimal()

p2 <- ggplot(customer_data, aes(y = Age)) +
  geom_boxplot(fill = "#349", alpha = 0.7) +
  labs(title = "Boxplot Umur Pelanggan", y = "Umur") +
  theme_minimal()

p1 + p2

kable(age_stats, caption = "Ukuran Pemusatan Data Umur") %>%
  kable_styling(bootstrap_options = c("striped", "hover"), full_width = FALSE)

purchase_stats <- customer_data %>%
  summarise(
    Mean = round(mean(TotalPurchase), 1),
    Median = median(TotalPurchase),
    Mode = as.numeric(modeest::mfv(TotalPurchase)[1]),
    SD = round(sd(TotalPurchase), 1),
    Min = min(TotalPurchase),
    Max = max(TotalPurchase)
  )

# Visualisasi total pembelian
p3 <- ggplot(customer_data, aes(x = TotalPurchase)) +
  geom_histogram(aes(y = after_stat(density)), bins = 20, fill = "#3498db", alpha = 0.7) +
  geom_density(linewidth = 1.1, color = "#2c3e50") +
  geom_vline(xintercept = purchase_stats$Mean, color = "#e74c3c", linewidth = 1, linetype = "dashed") +
  geom_vline(xintercept = purchase_stats$Median, color = "#27ae60", linewidth = 1, linetype = "dashed") +
  labs(
    title = "Distribusi Total Purchase",
    subtitle = paste0("Mean = ", round(purchase_stats$Mean,2),
                      " | Median = ", round(purchase_stats$Median,2),
                      " | Mode = ", purchase_stats$Mode),
    x = "Total Purchase",
    y = "Density"
  ) +
  theme_minimal(base_size = 13)

# P4: Boxplot untuk Total Purchase 
p4_box <- ggplot(customer_data, aes(y = TotalPurchase)) +
  geom_boxplot(fill = "#3498db", alpha = 0.7) +
  labs(title = "Boxplot Total Purchase", y = "Total Purchase") +
  theme_minimal(base_size = 13)

p3 + p4_box

kable(purchase_stats, caption = "Ukuran Pemusatan Data Total Pembelian") %>%
  kable_styling(bootstrap_options = c("striped", "hover"), full_width = FALSE)

visits_stats <- customer_data %>%
  summarise(
    Mean = round(mean(NumberOfVisits), 1),
    Median = median(NumberOfVisits),
    Mode = as.numeric(modeest::mfv(NumberOfVisits)[1]),
    SD = round(sd(NumberOfVisits), 1),
    Min = min(NumberOfVisits),
    Max = max(NumberOfVisits)
  )

# Visualisasi jumlah kunjungan
p5 <- ggplot(customer_data, aes(x = NumberOfVisits)) +
  geom_bar(aes(y = after_stat(count)), fill = "#e67e22", alpha = 0.7, color = "black") +
  geom_vline(xintercept = visits_stats$Mean, color = "#e74c3c", linewidth = 1, linetype = "dashed") +
  geom_vline(xintercept = visits_stats$Median, color = "#2ecc71", linewidth = 1, linetype = "dashed") +
   annotate("label",
           x = visits_stats$Mean,
           y = max(table(customer_data$NumberOfVisits)) * 0.8,  
           label = "Mean",
           color = "#e74c3c",
           fill = "white",
           alpha = 0.6,
           angle = 90,
           vjust = 1.25,
           hjust = -0.5,
           size = 4.5,
           label.size = 0) +
  annotate("label",
           x = visits_stats$Median,
           y = max(table(customer_data$NumberOfVisits)) * 0.6,  
           label = "Median",
           color = "#2ecc71",
           fill = "white",
           alpha = 0.6,
           angle = 90,
           vjust = -0.25,
           hjust = 1.5,
           size = 4.5,
           label.size = 0) +
  labs(title = "Distribusi Jumlah Kunjungan",
       x = "Jumlah Kunjungan", y = "Frekuensi") +
  theme_minimal()


p6 <- ggplot(customer_data, aes(y = NumberOfVisits)) +
  geom_boxplot(fill = "#e67e22", alpha = 0.7) +
  labs(title = "Boxplot Jumlah Kunjungan", y = "Jumlah Kunjungan") +
  theme_minimal()

p5 + p6

kable(visits_stats, caption = "Ukuran Pemusatan Data Jumlah Kunjungan") %>%
  kable_styling(bootstrap_options = c("striped", "hover"), full_width = FALSE)

feedback_stats <- customer_data %>%
  summarise(
    Mean = round(mean(FeedbackScore), 2),
    Median = median(FeedbackScore),
    Mode = as.numeric(modeest::mfv(FeedbackScore)[1]),
    SD = round(sd(FeedbackScore), 2),
    Min = min(FeedbackScore),
    Max = max(FeedbackScore)
  )

# Visualisasi feedback score 

p7 <- ggplot(customer_data, aes(x = FeedbackScore)) +
  geom_bar(aes(y = after_stat(prop), group = 1), 
           fill = "#1abc9c", alpha = 0.7, color = "black") +
  geom_vline(xintercept = feedback_stats$Mean,
             color = "#e74c3c", linewidth = 1, linetype = "dashed") +
  geom_vline(xintercept = feedback_stats$Median,
             color = "#2ecc71", linewidth = 1, linetype = "dashed") +
  annotate("label",
           x = feedback_stats$Mean,
           y = 0.22,
           label = "Mean",
           color = "#e74c3c",
           fill = "white",
           alpha = 0.6,
           angle = 90,
           vjust = -0.2,
           hjust = -0.5,
           size = 4.5,
           label.size = 0) +
  annotate("label",
           x = feedback_stats$Median,
           y = 0.22,
           label = "Median",
           color = "#2ecc71",
           fill = "white",
           alpha = 0.6,
           angle = 90,
           vjust = 1.3,
           hjust = -0.5,
           size = 4.5,
           label.size = 0) +
  labs(
    title = "Distribusi Feedback Score",
    x = "Feedback Score", y = "Proporsi"
  ) +
  theme_minimal() +
  coord_cartesian(clip = "off")

p8 <- ggplot(customer_data, aes(y = FeedbackScore)) +
  geom_boxplot(fill = "#1abc9c", alpha = 0.7) +
  labs(title = "Boxplot Feedback Score", y = "Feedback Score") +
  theme_minimal()

# Gabungkan kedua plot
p7 + p8  

kable(feedback_stats, caption = "Ukuran Pemusatan Data Feedback Score") %>%
  kable_styling(bootstrap_options = c("striped", "hover"), full_width = FALSE)

gender_stats <- customer_data %>%
  group_by(Gender) %>%
  summarise(
    Count = n(),
    Percentage = paste0(round(n()/nrow(customer_data)*100, 1), "%"),
    Mean_Age = round(mean(Age), 1),
    Median_Age = median(Age),
    Mean_Purchase = round(mean(TotalPurchase), 1),
    Median_Purchase = median(TotalPurchase),
    Mean_Visits = round(mean(NumberOfVisits), 1),
    Median_Visits = median(NumberOfVisits),
    Mean_Feedback = round(mean(FeedbackScore), 2)
  )

# Plot A: Mean Purchase (Bar Chart)  
p_gender_mean <- gender_stats %>%
  ggplot(aes(x = Gender, y = Mean_Purchase, fill = Gender)) +
  geom_bar(stat = "identity", position = "dodge", alpha = 0.8) +
  geom_text(aes(label = scales::dollar(Mean_Purchase)), vjust = -0.5, size = 3.5, fontface = "bold") +
  labs(title = "Rata-rata Pembelian", x = "Gender", y = "Rata-rata Pembelian (USD)") +
  theme_minimal(base_size = 12) +
  theme(legend.position = "none") +
  scale_y_continuous(labels = scales::dollar_format())

# Plot B: Distribusi Total Pembelian (Boxplot) 
p_gender_box <- customer_data %>%
  ggplot(aes(x = Gender, y = TotalPurchase, fill = Gender)) + #
  geom_boxplot(alpha = 0.7) +
  labs(title = "Distribusi Pembelian", x = "Gender", y = "Total Pembelian (USD)") + 
  theme_minimal(base_size = 12) +
  theme(legend.position = "none") +
  scale_y_continuous(labels = scales::dollar_format())

# Gabungkan kedua visualisasi
print(p_gender_mean + p_gender_box)

kable(gender_stats, caption = "Ukuran Pemusatan berdasarkan Gender") %>%
  kable_styling(bootstrap_options = c("striped", "hover"))

location_stats <- customer_data %>%
  group_by(StoreLocation) %>%
  summarise(
    Count = n(),
    Percentage = paste0(round(n()/nrow(customer_data)*100, 1), "%"),
    Mean_Age = round(mean(Age), 1),
    Median_Age = median(Age),
    Mean_Purchase = round(mean(TotalPurchase), 1),
    Median_Purchase = median(TotalPurchase),
    Mean_Visits = round(mean(NumberOfVisits), 1),
    Median_Visits = median(NumberOfVisits),
    Mean_Feedback = round(mean(FeedbackScore), 2)
  )

# Plot A: Mean Purchase (Bar Chart) 
p_location_mean <- location_stats %>%
  ggplot(aes(x = StoreLocation, y = Mean_Purchase, fill = StoreLocation)) +
  geom_bar(stat = "identity", position = "dodge", alpha = 0.8) +
  geom_text(aes(label = scales::dollar(Mean_Purchase)), vjust = -0.5, size = 3.5, fontface = "bold") +
  labs(title = "Rata-rata Pembelian", x = "Lokasi Toko", y = "Rata-rata Pembelian (USD)") +
  theme_minimal(base_size = 12) +
  theme(legend.position = "none") +
  scale_y_continuous(labels = scales::dollar_format())

# Plot B: Distribusi Total Pembelian (Boxplot)
p_location_box <- customer_data %>%
  ggplot(aes(x = StoreLocation, y = TotalPurchase, fill = StoreLocation)) + 
  geom_boxplot(alpha = 0.7) +
  labs(title = "Distribusi Pembelian", x = "Lokasi Toko", y = "Total Pembelian (USD)") + 
  theme_minimal(base_size = 12) +
  theme(legend.position = "none") +
  scale_y_continuous(labels = scales::dollar_format())

# Gabungkan kedua visualisasi
print(p_location_mean + p_location_box)

kable(location_stats, caption = "Ukuran Pemusatan berdasarkan Lokasi Toko") %>%
  kable_styling(bootstrap_options = c("striped", "hover"))

product_stats <- customer_data %>%
  group_by(ProductCategory) %>%
  summarise(
    Count = n(),
    Percentage = paste0(round(n()/nrow(customer_data)*100, 1), "%"),
    Mean_Age = round(mean(Age), 1),
    Median_Age = median(Age),
    Mean_Purchase = round(mean(TotalPurchase), 1),
    Median_Purchase = median(TotalPurchase),
    Mode_Purchase = mlv(TotalPurchase, method = "mfv"), 
    Mean_Visits = round(mean(NumberOfVisits), 1),
    Median_Visits = median(NumberOfVisits),
    Mean_Feedback = round(mean(FeedbackScore), 2)
  )



# Plot A: Mean Purchase 

p_product_mean <- product_stats %>%
  ggplot(aes(x = reorder(ProductCategory, Mean_Purchase),
             y = Mean_Purchase,
             fill = ProductCategory)) +
  geom_col(alpha = 0.8) +
  geom_text(aes(label = dollar(Mean_Purchase)),
            vjust = -0.3, size = 3.5, fontface = "bold") +
  labs(title = "Rata-rata Pembelian per Kategori Produk",
       x = "Kategori Produk",
       y = "Rata-rata Pembelian (USD)") +
  theme_minimal(base_size = 12) +
  theme(legend.position = "none",
        plot.title = element_text(face = "bold", hjust = 0.5)) +
  scale_y_continuous(labels = dollar_format())



# Plot B: Boxplot Distribusi Pembelian

p_product_box <- customer_data %>%
  left_join(product_stats %>% select(ProductCategory, Mean_Purchase),
            by = "ProductCategory") %>%
  ggplot(aes(x = reorder(ProductCategory, Mean_Purchase),
             y = TotalPurchase,
             fill = ProductCategory)) +
  geom_boxplot(alpha = 0.7) +
  labs(title = "Distribusi Total Pembelian",
       x = "Kategori Produk",
       y = "Total Pembelian (USD)") +
  theme_minimal(base_size = 12) +
  theme(legend.position = "none",
        plot.title = element_text(face = "bold", hjust = 0.5)) +
  scale_y_continuous(labels = dollar_format())



# Gabungan kedua visualisasi

print(p_product_mean / p_product_box)

# tabel ringkasan
kable(product_stats, caption = "Ukuran Pemusatan berdasarkan Kategori Produk") %>%
  kable_styling(bootstrap_options = c("striped", "hover"))