library(sf)
library(sp)


vars <- c("X1","X2","X3","X4","X5")



for (v in vars) {
  # Koefisien dan standar error
  coef_val <- gwr_sp@data[[v]]
  se_val <- gwr_sp@data[[paste0(v, "_se")]]
  t_val <- coef_val / se_val
  
  # 1. Simpan koefisien
  gwr_sp@data[[paste0("b_", v)]] <- coef_val
  
  # 2. Simpan status signifikan
  signif_flag <- ifelse(abs(t_val) > 1.96, "Ya", "Tidak")
  gwr_sp@data[[paste0("sig_", v)]] <- factor(signif_flag, levels = c("Tidak", "Ya"))
  
  # 3. Simpan koefisien hanya signifikan
  gwr_sp@data[[paste0("b_", v, "_sig")]] <- ifelse(abs(t_val) > 1.96, coef_val, NA)
  
  # --- Peta 1: Koefisien
  print(spplot(gwr_sp, zcol = paste0("b_", v),
               main = paste("Peta Koefisien", v),
               col.regions = colorRampPalette(c("blue", "white", "red"))(100),
               na.col = "grey"))
  
  # --- Peta 2: Signifikansi
  print(spplot(gwr_sp, zcol = paste0("sig_", v),
               main = paste("Signifikansi", v),
               col.regions = c("black", "red")))
  
  # --- Peta 3: Koefisien Hanya Signifikan
  print(spplot(gwr_sp, zcol = paste0("b_", v, "_sig"),
               main = paste("Koefisien Signifikan", v),
               col.regions = colorRampPalette(c("blue", "green", "red"))(100),
               na.col = "black"))
}

# Hitung jumlah kabupaten/kota yang signifikan (bernilai "Ya")
sig_counts <- sapply(vars, function(v) {
  kolom_nama <- paste0("sig_", v)
  sum(gwr_sp@data[[kolom_nama]] == "Ya", na.rm = TRUE)
})

# Buat data frame hasil
tabel7 <- data.frame(
  `Variabel Independen` = vars,
  `Jumlah Kabupaten/Kota Signifikan` = sig_counts
)

# Tampilkan tabel
kable(tabel7)

library(RColorBrewer)
library(classInt)
library(sp)

for (v in vars) {
  val <- gwr_sp@data[[paste0("b_", v, "_sig")]]
  
  # Inisialisasi vektor kategori NA
  cat_val <- rep(NA_character_, length(val))
  
  label_list <- c()
  color_list <- c()
  
  # --- NEGATIF ---
  val_neg <- val[val < 0 & !is.na(val)]
  if (length(val_neg) >= 3) {
    brk_neg <- classIntervals(val_neg, n = 3, style = "quantile")$brks
    neg_cuts <- cut(val, breaks = brk_neg, include.lowest = TRUE, dig.lab = 2)
    neg_lvls <- levels(neg_cuts)

    # Format label
    labels_neg <- gsub(",", ";", neg_lvls)
    labels_neg <- gsub("\\.", ",", labels_neg)
    
    cat_val[val < 0 & !is.na(val)] <- labels_neg[as.numeric(neg_cuts[val < 0 & !is.na(val)])]
    
    label_list <- c(label_list, labels_neg)
    color_list <- c(color_list, c("#8B0000", "#B22222", "#CD5C5C")[1:length(labels_neg)])
  }

  # --- POSITIF ---
  val_pos <- val[val > 0 & !is.na(val)]
  if (length(val_pos) >= 3) {
    brk_pos <- classIntervals(val_pos, n = 3, style = "quantile")$brks
    pos_cuts <- cut(val, breaks = brk_pos, include.lowest = TRUE, dig.lab = 2)
    pos_lvls <- levels(pos_cuts)

    # Format label
    labels_pos <- gsub(",", ";", pos_lvls)
    labels_pos <- gsub("\\.", ",", labels_pos)
    
    cat_val[val > 0 & !is.na(val)] <- labels_pos[as.numeric(pos_cuts[val > 0 & !is.na(val)])]
    
    label_list <- c(label_list, labels_pos)
    color_list <- c(color_list, c("#87CEFA", "#4682B4", "#00008B")[1:length(labels_pos)])
  }

  # --- Tidak Signifikan (NA) ---
  cat_val[is.na(cat_val)] <- "Tidak Signifikan"
  label_list <- c(label_list, "Tidak Signifikan")
  color_list <- c(color_list, "black")
  
  # Faktor berurutan
  cat_val <- factor(cat_val, levels = label_list)

  # Tambahkan ke data
  gwr_sp@data[[paste0("b_", v, "_sig_cat")]] <- cat_val

 

  # Plot
  print(
    spplot(gwr_sp, zcol = paste0("b_", v, "_sig_cat"),
           main = paste("Koefisien Signifikan", v),
           col.regions = setNames(color_list, label_list),
           par.settings = list(layout.heights = list(key.bottom = 2), axis.line = list(col = NA)),
           colorkey = list(
             space = "bottom",
             labels = list(
               at = 1:length(label_list),
               labels = label_list,
               cex = 1
             ),
             height = 1
           ))
  )

}

# Daftar variabel yang dianalisis (misalnya sudah didefinisikan sebelumnya)
vars <- c("X1","X2","X3","X4","X5")  # ganti sesuai variabelmu

# Inisialisasi hasil
hasil_tabel <- data.frame(Koef = c("Positif", "Negatif"), stringsAsFactors = FALSE)

# Loop per variabel (misalnya X1 sampai Xn)
for (v in vars) {
  kolom_nama <- paste0("b_", v, "_sig")  # Kolom berisi koefisien signifikan
  
  # Pastikan kolom ada
  if (!kolom_nama %in% names(gwr_sp@data)) next
  
  val <- gwr_sp@data[[kolom_nama]]
  nama <- gwr_sp@data$nmkab  # atau NAMOBJ jika itu nama wilayah
  
  # Filter yang signifikan dan positif/negatif
  positif <- nama[which(val > 0 & !is.na(val))]
  negatif <- nama[which(val < 0 & !is.na(val))]
  
  # Gabungkan nama kecamatan
  hasil_tabel[[v]] <- c(
    if (length(positif) > 0) paste(sort(positif), collapse = ", ") else "-",
    if (length(negatif) > 0) paste(sort(negatif), collapse = ", ") else "-"
  )
}

# Tampilkan hasil
kable(hasil_tabel, row.names = FALSE)

# Ambil kolom signifikan (TRUE/FALSE) dari variabel tertentu
sig_df <- gwr_sp@data[, paste0("sig_", vars)]

# Pastikan dalam bentuk logical TRUE/FALSE
sig_df <- lapply(sig_df, function(x) {
  if (is.factor(x)) as.logical(as.numeric(as.character(x) == "Ya")) else as.logical(x)
})
sig_df <- as.data.frame(sig_df)

# Buat kolom var_sig (nama variabel yang signifikan)
gwr_sf$var_sig <- apply(sig_df, 1, function(row) {
  signif_vars <- vars[which(row)]
  if (length(signif_vars) == 0) {
    return("Tidak ada")
  } else {
    return(paste(signif_vars, collapse = ", "))
  }
})

library(dplyr)

# --- Ambil nilai signifikan dari gwr_sp
sig_df <- gwr_sp@data[, paste0("sig_", vars)]

# --- Pastikan hasilnya logical
sig_df <- lapply(sig_df, function(x) {
  if (is.factor(x)) as.character(x) == "Ya" else as.logical(x)
})
sig_df <- as.data.frame(sig_df)

# --- Buat kolom var_sig langsung di @data
gwr_sp@data$var_sig <- apply(sig_df, 1, function(row) {
  signif_vars <- vars[which(row)]
  if (length(signif_vars) == 0) {
    return("Tidak ada")
  } else {
    return(paste(sort(signif_vars), collapse = ", "))
  }
})

# --- Rekap: nama kecamatan per kategori var_sig
tabel_kec_per_kategori <- gwr_sp@data %>%
  group_by(var_sig) %>%
  summarise(
    daftar_kabupaten = paste(sort(unique(nmkab)), collapse = ", "),
    jumlah = dplyr::n()
  ) %>%
  arrange(var_sig)

# --- Tampilkan semua hasil
kable(tabel_kec_per_kategori, n = Inf)

library(ggplot2)
library(sf)

# Ubah ke sf jika belum
gwr_sf <- st_as_sf(gwr_sp)
gwr_sf$var_sig <- factor(gwr_sf$var_sig)

warna_kategori_final <- c(
  "black", "#0000CD", "#4169E1", "#4682B4", "#5F9EA0", "#6495ED", "#87CEEB", "#87CEFA",
  "#8B0000", "#A52A2A", "#B22222", "#CD5C5C", "#DC143C", "#E9967A", "#FA8072", "#FFA07A",
  "#008000", "#32CD32", "#3CB371", "#00FA9A", "#20B2AA", "#40E0D0"
)

# Buat warna sesuai urutan faktor
warna_kategori_named <- setNames(warna_kategori_final, levels(gwr_sf$var_sig))

# Plot
ggplot(gwr_sf) +
  geom_sf(aes(fill = var_sig), color = "black", size = 0.2) +
  scale_fill_manual(
    values = warna_kategori_named,
    name = "Signifikansi Variabel Lokal",
    guide = guide_legend(
      direction = "horizontal",
      title.position = "top",
      label.position = "bottom",
      nrow = 2
    )
  ) +
  labs(title = "Peta Choropleth: Kelompok Kombinasi Variabel Signifikan (GWR)") +
  coord_sf(datum = NA) +  # Hilangkan label koordinat dan garis sumbu
  theme_void() +          # Hilangkan seluruh grid & axis
  theme(
    legend.position = "bottom",
    legend.title.align = 0.5,
    legend.text = element_text(size = 12),
    plot.title = element_text(hjust = 0.5, face = "bold")
  )