根据属性对相邻边进行图形化

时间:2018-12-01 18:58:36

标签: r igraph

对于每个顶点,我对基于条件的其相邻边的数量感兴趣。在以下示例中,条件具有不同的性别。

示例:

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


nodes <- tibble(id = 1:4, 
                gender = c("M", "F", "F", "M"), 
                names = c("Bob", "Allie", "Mary", "Johnathon"))

edges <- tibble(from = c(1, 3, 2, 4, 1, 2, 1, 4),
                to = c(2, 2, 4, 1, 3, 1, 4, 3))

network <- graph_from_data_frame(d = edges, vertices = nodes, directed = TRUE)

ggraph(network) + 
geom_edge_link(arrow = arrow(length = unit(4, 
'mm')), 
         start_cap = circle(4, 'mm'), 
         end_cap = circle(4, 'mm')) + 
geom_node_text(aes(label = names)) +
theme_graph()

Example Plot

所需结果:

id  name          adjacent_edges

1    Bob          1
2    Allie        1
3    Mary         2
4    Johnathon    1

1 个答案:

答案 0 :(得分:2)

这是将基数R与igraph组合在一起的一种方法:

nodes %>% 
  mutate(adjacent_edges = colSums(as.matrix(outer(gender, gender, `!=`) * as_adj(network)) != 0))
# A tibble: 4 x 4
#      id gender names     adjacent_edges
#   <int> <chr>  <chr>              <dbl>
# 1     1 M      Bob                    1
# 2     2 F      Allie                  1
# 3     3 F      Mary                   2
# 4     4 M      Johnathon              1

这里

outer(gender, gender, `!=`)

当性别不同时,用TRUE个条目构建一个矩阵,而as_adj(network))是通常的图邻接矩阵。然后,在具有不同性别的连接节点的情况下,恰好在需要时,它们的乘积将具有非零条目。对这些情况进行总结可以得出预期的结果。

这是另一个,更长,但也更透明:

edges %>% full_join(nodes, by = c("from" = "id")) %>% 
  full_join(nodes, by = c("to" = "id"), suff = c(".from", ".to")) %>%
  group_by(to, names.to) %>% summarise(adjacent_edges = sum(gender.to != gender.from)) %>%
  rename(id = to, name = names.to)
# A tibble: 4 x 3
# Groups:   id [4]
#      id name      adjacent_edges
#   <dbl> <chr>              <int>
# 1     1 Bob                    1
# 2     2 Allie                  1
# 3     3 Mary                   2
# 4     4 Johnathon              1

在这种情况下,我们从边列表开始,然后两次添加节点列表:一次是获取有关from边的节点信息,一次是获取有关{{1 }}边缘,在同一行中。然后剩下的是通过汇总所有不同性别的邻居来汇总数据。