如何让R语言visNetwork层级图正确展示数据中的层级关系？

问题描述 #

使用R的visNetwork绘制组织架构网络图时，已使用visHierarchicalLayout(direction = "UD")，但布局未正确反映层级关系，期望层级为：Boss（最顶部）→ Teams → Subteams → Employees。现有数据集和绘图代码如下：

数据集代码 #

library(igraph)
library(visNetwork)

nodes_df = structure(list(id = c("Boss", "TeamA", "TeamB", "SubteamA1", 
"SubteamA2", "SubteamA3", "SubteamB1", "SubteamB2", "SubteamB3", 
"employee1", "employee2", "employee3", "employee4", "employee5", 
"employee6", "employee7", "employee8", "employee9", "employee10", 
"employee11", "employee12", "employee13", "employee14", "employee15", 
"employee16", "employee17", "employee18"), label = c("Boss", 
"TeamA", "TeamB", "SubteamA1", "SubteamA2", "SubteamA3", "SubteamB1", 
"SubteamB2", "SubteamB3", "employee1", "employee2", "employee3", 
"employee4", "employee5", "employee6", "employee7", "employee8", 
"employee9", "employee10", "employee11", "employee12", "employee13", 
"employee14", "employee15", "employee16", "employee17", "employee18"
), group = c("yellow", "red", "red", "green", "green", "green", 
"green", "green", "green", "purple", "purple", "purple", "purple", 
"purple", "purple", "purple", "purple", "purple", "purple", "purple", 
"purple", "purple", "purple", "purple", "purple", "purple", "purple"
)), row.names = c(NA, -27L), class = "data.frame")

edges_df = structure(list(from = c("Boss", "TeamA", "TeamA", "TeamA", "SubteamA1", 
"SubteamA1", "SubteamA1", "SubteamA2", "SubteamA2", "SubteamA2", 
"SubteamA3", "SubteamA3", "SubteamA3", "Boss", "TeamB", "TeamB", 
"TeamB", "SubteamB1", "SubteamB1", "SubteamB1", "SubteamB2", 
"SubteamB2", "SubteamB2", "SubteamB3", "SubteamB3", "SubteamB3"
), to = c("TeamA", "SubteamA1", "SubteamA2", "SubteamA3", "employee1", 
"employee2", "employee3", "employee4", "employee5", "employee6", 
"employee7", "employee8", "employee9", "TeamB", "SubteamB1", 
"SubteamB2", "SubteamB3", "employee10", "employee11", "employee12", 
"employee13", "employee14", "employee15", "employee16", "employee17", 
"employee18")), row.names = c(NA, -26L), class = "data.frame")

原绘图代码 #

# Create the visNetwork object
visNetwork(nodes_df, edges_df) %>%   visHierarchicalLayout(direction = "UD") %>%
    visOptions(highlightNearest = TRUE, nodesIdSelection = TRUE) %>%
    visInteraction(navigation = "zoom") %>%
    visInteraction(navigation = "drag")

解决方案 #

默认层级布局算法没有明确的层级标记，需手动给每个节点指定层级，再启用基于层级的布局规则：

1. 给节点添加层级列：在nodes_df中新增level列，明确每个节点的层级：

   • Boss → 层级1（最顶部）
   • TeamA、TeamB → 层级2
   • 所有Subteam → 层级3
   • 所有employee → 层级4

2. 修改布局参数：在visHierarchicalLayout中设置levelBased = TRUE，让布局严格按照指定的层级排列。

修改后的完整代码 #

library(igraph)
library(visNetwork)
library(dplyr) # 需要加载dplyr以使用case_when

# 给nodes_df添加level列
nodes_df$level <- case_when(
  nodes_df$id == "Boss" ~ 1,
  nodes_df$id %in% c("TeamA", "TeamB") ~ 2,
  grepl("Subteam", nodes_df$id) ~ 3,
  grepl("employee", nodes_df$id) ~ 4
)

# 绘制网络图
visNetwork(nodes_df, edges_df) %>%
  visHierarchicalLayout(direction = "UD", levelBased = TRUE) %>%
  visOptions(highlightNearest = TRUE, nodesIdSelection = TRUE) %>%
  visInteraction(navigation = c("zoom", "drag"))

说明 #

• case_when用于根据节点id匹配对应层级，也可手动给每行指定level值，效果一致。
• levelBased = TRUE会强制visNetwork按照level列的值从上到下排列节点，确保层级关系完全符合预期。
• 将两个visInteraction合并为一个，用c("zoom", "drag")同时启用缩放和拖拽功能，代码更简洁。

内容的提问来源于stack exchange，提问作者stats_noob