为堆叠条形图的每个分栏设置专属父色及内部渐变色

Got it, let's break this down so you can get exactly the chart you want. You have 18 main categories (each bar on the x-axis), each stacked with 39 sub-levels, and you want every main bar to have its own unique parent color—with the sub-levels inside each bar fading from that parent color (light to dark, or reverse). Here's how to pull this off:

Step 1: Set Up Your Data (or Simulate It) #

First, let's assume your data looks like this (replace this with your actual dataset):

library(ggplot2)
library(scales)
library(dplyr)

# Simulate a dataset matching your structure: 18 main cats, 39 sub cats each
set.seed(123)
df <- expand.grid(
  main_category = paste0("Group_", 1:18),
  sub_level = paste0("Level_", 1:39)
) %>%
  mutate(value = sample(1:100, n(), replace = TRUE)) # Replace with your actual values

Step 2: Assign Parent Colors to Each Main Category #

We'll first pick 18 distinct parent colors (you can customize these to fit your brand or preferences):

# Generate 18 distinct parent colors using the hue palette (adjustable!)
parent_colors <- hue_pal(h = c(0, 360), c = 80, l = 60)(18)
names(parent_colors) <- unique(df$main_category)

Tweak the h, c, and l arguments in hue_pal() to adjust the hue range, saturation, and lightness of your parent colors if you want more contrast.

Step 3: Generate Gradient Colors for Each Sub-Level #

Next, we'll create a gradient of 39 colors for each main category, based on its parent color. We'll use colorRampPalette to fade from a lighter version of the parent color to the parent itself (reverse the order if you want dark-to-light):

# Create a list of gradient color sets (one per main category)
gradient_sets <- lapply(parent_colors, function(base_color) {
  # Fade from a lighter shade of the parent color to the parent color
  colorRampPalette(c(lighten(base_color, 0.5), base_color))(39)
})

# Map each sub-level in its main category to the correct gradient color
df <- df %>%
  group_by(main_category) %>%
  mutate(
    # Ensure sub-levels are ordered correctly (adjust if your sub_level has a natural order)
    sub_order = as.integer(factor(sub_level)),
    fill_color = gradient_sets[[unique(main_category)]][sub_order]
  ) %>%
  ungroup()

If your sub_level variable has a specific order (like numeric levels), use that instead of factor(sub_level) to make sure the gradient aligns with your hierarchy.

Step 4: Build the Plot #

Now we'll plot using the custom colors we generated. We use scale_fill_identity because we're mapping pre-defined colors directly:

ggplot(df, aes(x = main_category, y = value, fill = fill_color)) +
  geom_bar(stat = "identity", color = "grey", size = 0.2) + # Add grey borders for clarity
  # Use our custom colors and add a legend for sub-levels
  scale_fill_identity(
    guide = "legend",
    labels = unique(df$sub_level),
    name = "Sub Levels"
  ) +
  labs(x = "Main Categories", y = "Value") +
  # Rotate x-axis labels so they're readable
  theme(axis.text.x = element_text(angle = 45, hjust = 1, vjust = 1))

Quick Adjustments You Might Want #

• Reverse the gradient: Swap the order inside colorRampPalette to c(base_color, lighten(base_color, 0.5)) for dark-to-light fading.
• More distinct parent colors: Try using viridis_pal()(18) instead of hue_pal() for colorblind-friendly parent colors.
• Smoother gradients: If 39 steps feel abrupt, adjust the lightness value (e.g., lighten(base_color, 0.3) for a smaller range between light and dark).

内容的提问来源于stack exchange，提问作者nicole