如何在R中制作不同半径、可区分类别的弧图？

Absolutely! You can absolutely create arc diagrams where different connection categories are represented by arcs of varying radii—this is totally achievable in R, even if packages like ggraph or arcdiagram don’t support it out of the box. Let’s walk through two reliable approaches:

Approach 1: Layered Arcs with ggraph #

The trick here is to split your edges by category and draw each group with a custom radius using geom_edge_arc. Here’s a complete example:

Step 1: Prepare Data #

library(tidyverse)
library(ggraph)
library(igraph)

# Create edge data with category labels
edges <- tibble(
  from = c("A", "A", "B", "B", "C", "C"),
  to = c("D", "D", "E", "E", "F", "F"),
  category = c("Primary", "Secondary", "Primary", "Secondary", "Primary", "Secondary")
)

# Convert to an igraph object
graph <- graph_from_data_frame(edges, directed = FALSE)

Step 2: Assign Radii to Categories #

Map each category to a unique radius value (adjust these numbers based on how spread out you want the arcs):

radius_mapping <- tibble(
  category = c("Primary", "Secondary"),
  radius = c(0.25, 0.5)
)

# Merge radius values into edge data
edges_with_radius <- edges %>% left_join(radius_mapping, by = "category")

Step 3: Plot Layered Arcs #

Draw each category’s arcs separately with their assigned radius:

ggraph(graph, layout = "linear") +
  # Draw nodes
  geom_node_point(size = 6, color = "#2c3e50") +
  geom_node_text(aes(label = name), vjust = -1.2, size = 4) +
  
  # Draw arcs for each category
  geom_edge_arc(
    data = function(x) filter(x, category == "Primary"),
    aes(color = category),
    radius = 0.25,
    linewidth = 1.8
  ) +
  geom_edge_arc(
    data = function(x) filter(x, category == "Secondary"),
    aes(color = category),
    radius = 0.5,
    linewidth = 1.8
  ) +
  
  # Customize colors and theme
  scale_color_manual(values = c("Primary" = "#e74c3c", "Secondary" = "#3498db")) +
  theme_void()

This will give you arcs with distinct radii for each category, no overlapping!

Approach 2: Manual Arc Drawing with ggforce #

For even more control, use ggforce::geom_arc to calculate and draw arcs from scratch. This is great if you need precise control over arc positioning:

Step 1: Get Node Positions #

First, generate linear layout positions for your nodes:

library(ggforce)

# Create linear layout and extract node positions
layout <- create_layout(graph, layout = "linear")
node_pos <- layout %>% select(name, x)

Step 2: Calculate Arc Parameters #

Compute the center, radius, and angles for each arc based on its category:

arc_params <- edges_with_radius %>%
  left_join(node_pos, by = c("from" = "name")) %>%
  rename(x_start = x) %>%
  left_join(node_pos, by = c("to" = "name")) %>%
  rename(x_end = x) %>%
  mutate(
    x_center = (x_start + x_end) / 2,  # Midpoint between nodes
    # Set arc direction (upwards)
    start_angle = ifelse(x_start < x_end, pi, 0),
    end_angle = ifelse(x_start < x_end, 0, pi)
  )

Step 3: Plot Custom Arcs #

Use geom_arc to draw each arc with its calculated parameters:

ggplot() +
  # Draw nodes
  geom_point(data = node_pos, aes(x = x, y = 0), size = 6, color = "#2c3e50") +
  geom_text(data = node_pos, aes(x = x, y = 0.6, label = name), size = 4) +
  
  # Draw custom arcs
  geom_arc(
    data = arc_params,
    aes(
      x0 = x_center, y0 = 0,
      r = radius,
      start = start_angle, end = end_angle,
      color = category
    ),
    linewidth = 1.8
  ) +
  
  # Customize plot
  scale_color_manual(values = c("Primary" = "#e74c3c", "Secondary" = "#3498db")) +
  coord_fixed() +
  theme_void()

Key Notes #

• Adjust the radius values to control how high each arc curves—larger values mean taller arcs.
• For directed graphs, you can tweak the start_angle and end_angle parameters to make arcs curve in different directions (e.g., upwards for forward connections, downwards for reverse).
• Both methods ensure arcs from different categories don’t overlap, making your diagram much easier to interpret.

内容的提问来源于stack exchange，提问作者santoku