How to compute average closeness of ego subnetwork based on alter attributes in R?

Question

I have ego-network data. Participants categorized 3 of their friends as A, B, or C. Then, they rated how close they are to one another.

I want to create a column with the average closeness between each participant's friends categorized as type "C".

For example, for the first row, only "friend1_closeness_friend2" would be used in the calculation and "friend1_closeness_friend3" and "friend2_closeness_friend3" would be ignored.

library(dplyr)

data <- data.frame(
  ID = c("001", "002", "003"),
  friendshipType_1 = c("C", "C", "A"),
  friendshipType_2 = c("C", "C", "B"),
  friendshipType_3 = c("A", "C" , "A"),
  friend1_closeness_friend2 = c(1, 2, 3),
  friend1_closeness_friend3 = c(4, 3, 2),
  friend2_closeness_friend3 = c(1, 1, 2)
)

Desired output:

data <- data.frame(
  ID = c("001", "002", "003"),
  friendshipType_1 = c("C", "C", "A"),
  friendshipType_2 = c("C", "C", "B"),
  friendshipType_3 = c("A", "C" , "A"),
  friend1_closeness_friend2 = c(1, 2, 3),
  friend1_closeness_friend3 = c(4, 3, 2),
  friend2_closeness_friend3 = c(1, 1, 2),
  mean_c = c(1, 2, NA)
)

Answer 1

Maybe you can achieve this using dplyr and rowwise() operation to calculate the mean closeness for each row based on the friendship type specified. Here's how you can do it:

  data2 <- data %>%
  rowwise() %>%
  mutate(mean_c = mean(c(friend1_closeness_friend2, friend1_closeness_friend3, friend2_closeness_friend3)[c(friendshipType_1, friendshipType_2, friendshipType_3) == "C"], na.rm = TRUE))

the mean_c column represents the mean closeness among friends categorized as type "C" for each participant. Hope it helps you!