如何在R的sunburstR包中指定分类颜色与标签及理解相关参数？

Great questions about sunburstR's sunburst() function—let's break down each part clearly, since the docs can be a bit vague on these details!

How Color Vectors Are Applied to Categories #

When you pass a simple color vector (like colors = c("#ff0000", "#00ff00", "#0000ff")) to sunburst(), the function uses D3's ordinal color scaling under the hood. Here's the exact flow:

• It extracts all unique node values from your hierarchical data (from root to every leaf node).
• It maps these unique nodes to your color vector in the order the nodes first appear in your data (or sorted order, depending on your data structure).
• If your color vector is shorter than the number of unique nodes, it will cycle through the colors repeatedly.

For example, if your data has root node "Total", first-level nodes "A" / "B", and second-level nodes "A1" / "A2" / "B1", a vector c("red", "blue") would assign:

• "Total" → red
• "A" → blue
• "B" → red (cycles back)
• "A1" → blue
• "A2" → red
• "B1" → blue

Manual Color Assignment for Root/Leaf Nodes (Better Approaches) #

A simple color vector works for general cases, but for targeting root or leaf nodes specifically, using a custom color mapping list is far more reliable. Here's how to do it:

Targeting the Root Node #

If your root node has a fixed name (e.g., "Root"), define a list with domain (the nodes you want to target) and range (their corresponding colors):

root_color_map <- list(
  domain = c("Root"),
  range = c("#000000")  # Black root node
)
sunburst(your_data, colors = root_color_map)

Any nodes not in domain will use the default D3 categorical color scheme.

Targeting Leaf Nodes #

First, identify all leaf nodes from your data (this depends on your data format—for CSV-based hierarchies, leaves are the deepest non-NA values). Then add them to your color map:

# Example: Data has columns level1, level2, level3; leaves are level3 values
leaf_nodes <- unique(your_data$level3[!is.na(your_data$level3)])

leaf_color_map <- list(
  domain = c("Root", leaf_nodes),
  range = c("#000000", rep("#ff0000", length(leaf_nodes)))  # Black root, red leaves
)
sunburst(your_data, colors = leaf_color_map)

Pro Tip: Override All Node Colors #

If you want full control, include every unique node in domain and map each to a specific color in range.

By default, sunburst() displays labels for all nodes, but you can tweak this to avoid clutter or highlight specific partitions:

• Hide labels entirely: Set labels = FALSE.
• Show labels only for large partitions: Use labelThreshold to set a minimum percentage of the total. For example, labelThreshold = 0.05 will only show labels for partitions making up 5% or more of the total.
• Adjust label appearance: Use labelSize (font size) and labelAngle (rotation) to improve readability for nested partitions.

Example:

sunburst(your_data, labelThreshold = 0.03, labelSize = 12, labelAngle = 45)

This line in the docs refers to the flexible color mapping options when you pass a list to the colors parameter:

• Only range: If you provide just a range (e.g., list(range = c("red", "blue"))), the function automatically uses all unique nodes as the domain, mapping them to your range colors (cycling if needed). This is equivalent to passing a simple color vector, but more explicit.
• Only domain: If you provide just a domain (e.g., list(domain = c("Root", "A"))), the function uses the default D3 color scheme for those nodes, and falls back to defaults for others.
• Both domain and range: This is the most powerful option—you explicitly map specific node values to specific colors, which is exactly what you need for targeting root/leaf nodes.

内容的提问来源于stack exchange，提问作者Carmen Sandoval