igraph使用遇阻，如何编写循环批量处理个体网络分析统计？

需求 #

从包含数百个个体的地点检测数据集中，批量计算每个个体的三个网络统计量：

• 节点数：该个体访问的不同地点总数
• 边数：该个体在不同地点间的唯一连接数量（排除同地点的重复检测记录）
• 网络直径：该个体访问的任意两个地点之间的最大距离

数据集示例 #

生成简化数据集的代码 #

df <- data.frame(id = c("3811","3811","3832","3832","3832","3832"),
             Program = c("P1","P1","P1","P1","P1","P1"),
             from = c("hill","town","hill","wood","wood","lake"),
             from_lon = c(130.2,130.5,130.2,131.3,131.3,129.6),
             from_lat = c(-30.2,-30.5,-30.2,-31.3,-31.3,-29.6),
             to = c("town","lake","wood","wood","lake","town"),
             to_lon = c(130.5,129.6,131.3,131.3,129.6,130.5),
             to_lat = c(-30.5,-29.6,-31.3,-31.3,-29.6,-30.5),
             dist = c(44.111,132.506,161.456,0,249.847,132.506))

数据集内容 #

id Program from from_lon from_lat   to to_lon to_lat    dist
3811      P1 hill    130.2    -30.2 town  130.5  -30.5  44.111
3811      P1 town    130.5    -30.5 lake  129.6  -29.6 132.506
3832      P1 hill    130.2    -30.2 wood  131.3  -31.3 161.456
3832      P1 wood    131.3    -31.3 wood  131.3  -31.3   0.000
3832      P1 wood    131.3    -31.3 lake  129.6  -29.6 249.847
3832      P1 lake    129.6    -29.6 town  130.5  -30.5 132.506

当前手动实现 #

已编写代码可计算单个个体的统计量，但无法批量处理所有个体，代码如下：

# 计算个体3811的统计量
indiv_nodes <- df %>% 
  filter(id == "3811"& dist > 0) %>% # 排除同地点重复检测
  summarise(
    id = dplyr::first(id),
    prog = first(Program),
    nodes = n_distinct(to)+1, # +1包含起始地点
    netdiam = max(dist))
indiv_edges <- df %>% 
  filter(id == "3811" & dist > 0) %>% 
  group_by(from, to) %>%
  summarise( 
    from = dplyr::first(from),
    to = dplyr::first(to),
    weight = n())
net <- transform(indiv_nodes, edges = sum(indiv_edges$weight))

# 计算个体3832的统计量
indiv_nodes_n <- df %>% 
  filter(id == "3832" & dist > 0) %>% 
  summarise(
    id = dplyr::first(id),
    prog = first(Program),
    nodes = n_distinct(to)+1, 
    netdiam = max(dist))
indiv_edges_n <- df %>% 
  filter(id == "3832" & dist > 0) %>% 
  group_by(from, to) %>%
  summarise( 
    from = dplyr::first(from),
    to = dplyr::first(to),
    weight = n())
indiv_net <- transform(indiv_nodes_n, edges = sum(indiv_edges_n$weight))
net <- rbind(net, indiv_net)

# 输出结果
net

运行结果 #

id   prog  nodes   netdiam edges
3811   P1     3    132.506    2
3832   P1     4    249.847    3

求助点 #

数据集包含数百个个体，手动逐个处理效率极低，尝试编写循环函数未成功。希望得到以下任一解决方案：

1. 使用igraph包实现批量计算的代码
2. 将现有代码改造为可遍历所有id的批量处理函数

内容的提问来源于stack exchange，提问作者Boogaloo