如何根据阈值为ggraph边设置指定颜色？解决自定义颜色失效问题

你遇到的问题是：用ggraph绘制网络图时，想根据edges数据框的weight值（以50为阈值）区分边的颜色，但scale_color_manual设置的自定义颜色没生效，反而用了随机颜色。

原代码如下：

library(ggraph)
library(igraph)

# Data frame nodes
nodes <- data.frame(
  id = c(1, 2, 3, 4, 5),
  label = c("Especie1", "Especie2", "Especie3", "Especie4", "Especie5")
)

# Data frame edges
edges <- data.frame(
  from = c(1, 2, 3, 4, 5, 1, 3),
  to = c(2, 3, 4, 5, 1, 4, 5),
  weight = c(60, 40, 30, 70, 80, 90, 20)
)

# I create the graph
graph <- graph_from_data_frame(edges, vertices = nodes)

# Graphic with ggraph

ggraph(graph, layout = "stress") +
  geom_edge_link(aes(color = ifelse(edges$weight > 50, "Above Threshold", "Below Threshold"))) +
  geom_node_text(aes(label = label)) +
  scale_color_manual(values = c("Above Threshold" = "red", "Below Threshold" = "yellow")) +
  theme_void()

问题原因 #

你在geom_edge_link的aes里直接引用了外部的edges$weight，但ggraph在渲染图层时，是基于igraph对象内部的边属性，而非原始的edges数据框。这种外部引用会导致ggplot无法正确将这个衍生变量识别为映射的离散分类，进而忽略scale_color_manual的自定义设置。

解决方案 #

直接在aes中引用igraph对象自带的weight边属性即可，因为graph_from_data_frame已经将edges的所有列转化为图的边属性。

修正后的代码：

library(ggraph)
library(igraph)

nodes <- data.frame(
  id = c(1, 2, 3, 4, 5),
  label = c("Especie1", "Especie2", "Especie3", "Especie4", "Especie5")
)

edges <- data.frame(
  from = c(1, 2, 3, 4, 5, 1, 3),
  to = c(2, 3, 4, 5, 1, 4, 5),
  weight = c(60, 40, 30, 70, 80, 90, 20)
)

graph <- graph_from_data_frame(edges, vertices = nodes)

ggraph(graph, layout = "stress") +
  # 直接引用边属性weight，而非外部edges$weight
  geom_edge_link(aes(color = ifelse(weight > 50, "Above Threshold", "Below Threshold"))) +
  geom_node_text(aes(label = label)) +
  scale_color_manual(values = c("Above Threshold" = "red", "Below Threshold" = "yellow")) +
  theme_void()

另一种更规范的写法（提前创建分类变量） #

你也可以先给igraph对象的边添加一个分类属性，再在映射中使用，这样代码可读性更好：

# 给边添加分类属性
E(graph)$weight_category <- ifelse(E(graph)$weight > 50, "Above Threshold", "Below Threshold")

ggraph(graph, layout = "stress") +
  geom_edge_link(aes(color = weight_category)) +
  geom_node_text(aes(label = label)) +
  scale_color_manual(values = c("Above Threshold" = "red", "Below Threshold" = "yellow")) +
  theme_void()

这两种写法都能让ggplot正确识别离散分类变量，从而应用你设置的自定义颜色。

内容的提问来源于stack exchange，提问作者Jime Chaves