I have a set of ego-centric networks which I am analyzing using the egor package. The egor object is basically a list of three types of data named: ego, alter and aatie.
The ego data has one row of data per egoID. The alter data has multiple rows per egoID. Essentially, I'd like the ego data to have a column which counts the number of rows that exist for each egoID in the alter dataset. The alter data also has a column called "nodeID" which starts at 1 and goes sequentially up to n number of rows per egoID.
I'd like to remove all networks which have fewer than 10 alters. I thought I would start by creating a new column in the ego dataset called 'netsize' which has the number of alters for that ego. I can then filter from there. For the first step, I've tried several types of code, but to give you an idea of the mess I've made, here's one example:
egor.obj$ego <- egor.obj$ego %>%
mutate(netsize = map_dbl(egor.obj$alter, ~ max(.x$nodeID, na.rm = TRUE)))
I receive the following error:
Error in
mutate(): ℹ In argument:netsize = map_dbl(egor.obj$alter, ~max(.x$nodeID, na.rm = TRUE)). Caused by error inmap_dbl(): ℹ In index: 1. ℹ With name: .altID. Caused by error in.x$nodeID: ! $ operator is invalid for atomic vectors
I know egor can work with tidyverse but I'm quite new to both, and the egor object type is confusing me. What can I try next?
I have a set of ego-centric networks which I am analyzing using the egor package. The egor object is basically a list of three types of data named: ego, alter and aatie.
The ego data has one row of data per egoID. The alter data has multiple rows per egoID. Essentially, I'd like the ego data to have a column which counts the number of rows that exist for each egoID in the alter dataset. The alter data also has a column called "nodeID" which starts at 1 and goes sequentially up to n number of rows per egoID.
I'd like to remove all networks which have fewer than 10 alters. I thought I would start by creating a new column in the ego dataset called 'netsize' which has the number of alters for that ego. I can then filter from there. For the first step, I've tried several types of code, but to give you an idea of the mess I've made, here's one example:
egor.obj$ego <- egor.obj$ego %>%
mutate(netsize = map_dbl(egor.obj$alter, ~ max(.x$nodeID, na.rm = TRUE)))
I receive the following error:
Error in
mutate(): ℹ In argument:netsize = map_dbl(egor.obj$alter, ~max(.x$nodeID, na.rm = TRUE)). Caused by error inmap_dbl(): ℹ In index: 1. ℹ With name: .altID. Caused by error in.x$nodeID: ! $ operator is invalid for atomic vectors
I know egor can work with tidyverse but I'm quite new to both, and the egor object type is confusing me. What can I try next?
If I understand your data structure correctly, try something like this:
egor.obj$ego = egor.obj$ego %>%
left_join(egor.obj$alter %>%
count(egoID, name = "netsize"),
by = "egoID")
It counts the number of rows in alter per egoID returning a data frame with two columns egoID and netsize, then joins it with the ego data frame.
Another option with max nodeID per egoID:
egor.obj$ego = egor.obj$ego %>%
left_join(egor.obj$alter %>%
summarise(netsize = max(nodeID, na.rm = T), .by = "egoID"),
by = "egoID")
Check this article on how to filter and subset egor datasets.
This should work for your purpose (not tested):
egor.new = subset(egor.obj, subset = egor.obj$ego$netsize >= 10, unit="ego")
%>%) somewhat incorrectly (or at least non-traditionally) which, based on the error, may be contributing. You may find it useful to read how to properly use these pipes here, here, or here. As MrFlick said, it otherwise difficult to help with specifics without a reproducible example – jpsmith Commented Jan 2 at 18:50