可视化社交网络以显示用户在 R 中被提及的频率

Question

给定如下数据框：

v1    v2  v3     v4   v5
tom   A    pinky  A   3
ben   B    hugo   C   2
lily  A    tom    A   1
...

表示 v2 组的 v1 提到 v4 组的 v3 v5 次。例如，A 组的 tom 提到了 3 次 A 组的 pinky。现在我想绘制一个社交网络，每个用户用一个点表示，它的大小与他或她被完全提及的次数成正比。且两点相互或单边提及，则有一条连线。

当我查看 ggplot 文档时，我找不到任何函数来执行此操作。

你有什么想法吗？提前致谢！

编辑：

这是我目前得到的图表：

Answer 1

library(igraph)
library(dplyr)

# create example dataset
dt = data.frame(v1 = c("tom", "ben", "lilly", "mark"),
                v2 = c("A","B","A","C"),
                v3 = c("pinky", "hugo", "tom", "pinky"),
                v4 = c("A","D","A","A"),
                v5 = c(20,10,15,15),
                stringsAsFactors = F)

dt

#      v1 v2    v3 v4 v5
# 1   tom  A pinky  A 20
# 2   ben  B  hugo  D 10
# 3 lilly  A   tom  A 15
# 4  mark  C pinky  A 15


# select columns of names to use for the graph
dt_graph = dt %>% select(v1,v3)

# create the graph
g = graph.data.frame(dt_graph)

# count number of times names were mentioned
dt_times_mentioned =
  dt %>%
  group_by(v3) %>%
  summarise(times = sum(v5))

dt_times_mentioned

#      v3 times
#   (chr) (dbl)
# 1  hugo    10
# 2 pinky    35
# 3   tom    15


# join back to the vertex names to include names in the graph that were not mentioned
dt_weights =
  data.frame(names = names(V(g)), stringsAsFactors = F) %>%
  left_join(dt_times_mentioned, by=c("names"="v3")) %>%
  mutate(times = ifelse(is.na(times), 0, times))

dt_weights

#   names times
# 1   tom    15
# 2   ben     0
# 3 lilly     0
# 4  mark     0
# 5 pinky    35
# 6  hugo    10


# create two datasets based on 1st and 2nd column of names
dt1 = dt %>% select(names=v1, group=v2) 
dt2 = dt %>% select(names=v3, group=v4)


# get distinct names and their group values
dt_group = 
  dt1 %>% rbind(dt2) %>% distinct() %>% 
  mutate(color = colors()[as.numeric(factor(group))+5]) # get colours from group values

#(note that the +5 above is an arbitrary value for this example in order to get colors easy to distinguish. this is not really needed. if you don't have many groups you can manually set the colors)

dt_group

#   names group         color
# 1   tom     A antiquewhite3
# 2   ben     B antiquewhite4
# 3 lilly     A antiquewhite3
# 4  mark     C    aquamarine
# 5 pinky     A antiquewhite3
# 6  hugo     D   aquamarine1


# plot the graph
plot(g, vertex.size = dt_weights$times, vertex.color = dt_group$color)

# add legend
legend(1.5, 1.5,
       legend=unique(dt_group$group),
       pch=19,
       col=unique(dt_group$color),
       title = "Colors - Groups")

可视化社交网络以显示用户在 R 中被提及的频率

Visualize a social network to show how often a user is mentioned in R

r

data-visualization

social-networking

ggplot2