将 Igraph 转换为 VisNetwork

Question

我有使用“igraph”库制作的网络图：

library(tidyverse)
library(igraph)


set.seed(123)
n=15
data = data.frame(tibble(d = paste(1:n)))

relations = data.frame(tibble(
  from = sample(data$d),
  to = lead(from, default=from[1]),
))

data$name = c("new york", "chicago", "los angeles", "orlando", "houston", "seattle", "washington", "baltimore", "atlanta", "las vegas", "oakland", "phoenix", "kansas", "miami", "newark" )

graph = graph_from_data_frame(relations, directed=T, vertices = data) 
(edge_fac <- forcats::as_factor(get.edgelist(graph)[,1]))
n2 <- as.integer(factor(data$name,levels = levels(edge_fac)))


V(graph)$color <- ifelse(data$d == relations$from[1], "red", "orange")
V(graph)$label <- paste0(data$name,"\n\n\n",n2)
plot(graph, layout=layout.circle, edge.arrow.size = 0.2, main = "my_graph")

是否可以通过某种方式将上面的图转换为“visnetwork”图，使其看起来像这样？

我知道有一个函数 ( visIgraph() ) 用于将“igraph”图转换为“visnetwork”图：https://www.rdocumentation.org/packages/visNetwork/versions/2.1.0/topics/visNetwork-igraph

但我不确定是否可以将第一个“igraph”图（带有“数字”和“文本”标签）转换为交互式“visnetwork”图。

我尝试使用以下代码执行此操作：

visIgraph(graph)

但这会创建一个没有“数字”标签的交互式图表：

可以这样做吗？

谢谢！

Answer 1

使用 visNetwork 创建图形怎么样？然后，您可以在节点内添加编号和名称作为标签。

library(tidyverse)
library(visNetwork)

set.seed(123)
n=15
data = data.frame(tibble(id = paste(1:n)))

relations = data.frame(tibble(
  from = sample(data$id),
  to = lead(from, default=from[1]),
))

data$name = c("new york", "chicago", "los angeles", "orlando", "houston", "seattle", "washington", "baltimore", "atlanta", "las vegas", "oakland", "phoenix", "kansas", "miami", "newark" )
data$shape ='circle'
data$label = paste0(data$id,'\n',data$name)
data$color = ifelse(data$id==1, 'red', 'orange')

visNetwork(data, relations, width = "100%") %>%   
  visEdges(arrows =list(to = list(enabled = TRUE))) %>% 
  visIgraphLayout(layout = "layout_in_circle")

Answer 2

你必须做一些操作才能完成这项工作，因为它使用了基本的 R 绘图。

从本质上讲，这是两个不同的 igraph 对象，彼此重叠。这是我能想到的拥有两种不同 'cex' 尺寸的唯一方法。这可能需要一些技巧，具体取决于您从这里走向何方。

library(tidyverse)
library(igraph)
library(gridGraphics)    # <--- I'm new!
library(grid)            # <--- I'm new!

#----------- from question -----------
set.seed(123)
n=15
data = data.frame(tibble(d = paste(1:n)))    

relations = data.frame(tibble(
  from = sample(data$d),
  to = lead(from, default=from[1]),
))

data$name = c("new york", "chicago", "los angeles", "orlando",
              "houston", "seattle", "washington", "baltimore", 
              "atlanta", "las vegas", "oakland", "phoenix", 
              "kansas", "miami", "newark" )


graph = graph_from_data_frame(relations, 
                              directed=T, 
                              vertices = data) 
(edge_fac <- forcats::as_factor(get.edgelist(graph)[,1]))
n2 <- as.integer(factor(data$name,levels = levels(edge_fac)))
V(graph)$color <- ifelse(data$d == relations$from[1], 
                         "red", "orange")

这是变化的开始。

#---------- prepare the first plot -----------
# make label text larger
V(graph)$label.cex = 1.5
# V(graph)$label <- paste0(data$name,"\n",n2)
V(graph)$label <- paste0(n2) # just the number instead

#---------- prepare to collect grob ----------
# collect base plot grob
grabber <- function(){
  grid.echo()
  grid.grab()
}

# create a copy for the top layer
graph2 <- graph

#-------------- plot and grab ----------------
# without arrow sizes
plot(graph, layout=layout.circle, main = "my_graph")

# grab the grob
g1 = grabber()

现在是第二张图；顶层

#----------- create the top layer -------------
# with the copy, make the vertices transparent
V(graph2)$color <- "transparent"

# reset the font size
V(graph2)$label.cex = 1 

# shift the labels below (while keeping the plot design the same)
V(graph2)$label <- paste0("\n\n\n\n", data$name)

# show me
plot(graph2, layout=layout.circle,
     main = "my_graph", 
     edge.color = "transparent") # invisible arrows/ only 1 layer of arrows

# grab the grob
g2 = grabber()

将它们分层！

#-------------- redraw the plots -------------
# make the plot background transparent on the top layer
g2[["children"]][["graphics-background"]][["gp"]][["fill"]] <- "transparent"

# draw it!
grid.draw(g1)
grid.draw(g2)

您可能会发现有趣的是，进入 grob 的图表看起来与从中出来的不同...grid 本质上是对它们进行了调整。我觉得这有点棒。

将 Igraph 转换为 VisNetwork

Converting Igraph to VisNetwork

interactive

r

data-visualization

graph