How to get all the Twitter followers ids (> 75000) of a user using pagination in R with rtwitter library?

Question

I need a function to get all the followers (or friends) ids of a user in a data.frame with more than 75000 followers (or friends, above Twitter API limit) using the rtwitter library.

I first tried the next_cursor documentation help example:

# Retrieve user ids of accounts following POTUS
f1 <- get_followers("potus", n = 75000)
page <- next_cursor(f1)

# max. number of ids returned by one token is 75,000 every 15
# minutes, so you'll need to wait a bit before collecting the
# next batch of ids
sys.Sleep(15*60) # Suspend execution of R expressions for 15 mins

# Use the page value returned from \code{next_cursor} to continue
# where you left off.
f2 <- get_followers("potus", n = 75000, page = page)

But, how can I get all the followers (or friends) ids in the same data.frame taking into account different followers or friends counts for different users? How can I manage pagination?

smappR package take care of rate limits by itself: timeline, followers and friends without worries. Give it a try! — Tito Sanz, Apr 28 '18 at 00:39

Guz · Accepted Answer · 2017-07-17T22:13:35.250

After some test, I wrote this recursive function that include a progress bar and get all the followers ids from a user:

# Load library
library('rtweet')

# Global variables:
ids <- 75000 # Max number of ids per token every 15 minutes
f <- list() # Vector where user_ids from followers will be appended

# Function to get all the followers from a user with pagination
GetFollowersRecursivePagination <- function(userId, followers, page) {

  if (ids == 0) {

    # API Twitter Limit reached - Wait
    message("Waiting 15 mins...")
    total <- 15*60 # Total time = 15 min ~ 900 sec
    pb <- txtProgressBar(min = 0, max = total, style = 3) # create progress bar

    for (i in 1:total) {
      Sys.sleep(time = 1) # 1 second interval
      setTxtProgressBar(pb, i) # update progress bar
    }
    close(pb)

    # Check rate limit followers/ids query
    if (!rate_limit(token = NULL)[38,]$reset > 14.9) {
      message("Waiting 15 seconds more...")
      Sys.sleep(time = 15) # wait 15 seconds more...
    }

    message("Go!")
    ids <<- 75000
  }

  if (followers <= ids) {

    message(paste("Followers < ids | Number of Followers: ",
                  followers, " | Number of resting ids: ",  ids, sep = ""))
    ftemp <- get_followers(user = userId, n = followers, page = page)

    if (page == '-1') {
      f <<- append(f, list(ftemp)) # append followers ids
    }

    if (page != '-1') {
      df <- data.frame('user_id' = ftemp)
      f <<- append(f, list(df)) # append followers ids
    }

    ids <<- ids - followers
    message("Finished!")
    rtemp <- f
    f <<- list()
    return(rtemp)

  } else if (followers > ids) {

    message(paste("Followers > ids | Number of Followers: ",
                  followers, " | Number of resting ids: ",  ids, sep = ""))
    ftemp <- get_followers(user = userId, n = ids, page = page)

    if (page == '-1') {
      f <<- append(f, list(ftemp)) # append followers ids
    }

    if (page != '-1') {
      df <- data.frame('user_id' = ftemp)
      f <<- append(f, list(df)) # append followers ids
    }

    n <- ids # n = count of followers ids already acquired

    pageTemp <- next_cursor(ftemp) # Pagination

    # API Twitter Limit reached - Wait
    message("Waiting 15 mins...")
    total <- 15*60 # Total time = 15 min ~ 900 sec
    pb <- txtProgressBar(min = 0, max = total, style = 3) # create progress bar

    for (i in 1:total) {
      Sys.sleep(time = 1) # 1 second interval
      setTxtProgressBar(pb, i) # update progress bar
    }
    close(pb)

    # Check rate limit followers/ids query
    if (!rate_limit(token = NULL)[38,]$reset > 14.9) {
      message("Waiting 15 seconds more...")
      Sys.sleep(time = 15) # wait 15 seconds more...
    }

    message("Go!")
    ids <<- 75000

    # Recursive function call
    GetFollowersRecursivePagination(userId = userId,
                                    followers = followers - n,
                                    page = pageTemp)
  }
}

# Test ( > 75000 followers )
# Get user
user1 <- lookup_users(users = "146620155")
FAOClimate <- GetFollowersRecursivePagination(userId = user1$user_id,
                                              followers = user1$followers_count,
                                              page = '-1')

# Output:
Followers > ids | Number of Followers: 87208 | Number of resting ids:  75000
Followers < ids | Number of Followers: 12208 | Number of resting ids: 62792
Finished!

str(FAOClimate)

List of 2
 $ :'data.frame':   75000 obs. of  1 variable:
  ..$ user_id: chr [1:75000] "87189802" "884863013928546304" "886995444512964608" "852940633652301824" ...
  ..- attr(*, "next_cursor")= num 1.45e+18
 $ :'data.frame':   12207 obs. of  1 variable:
  ..$ user_id: chr [1:12207] "2175314977" "2168898233" "1491745484" "2175065456" ...

Note: change the function get_followers to get_friends if you want friends instead of followers.

score 2 · Answer 2 · answered Jul 25 '17 at 05:00

2

By default the get_followers function in the dev version of rtweet on Github can now be used to exceed 75k. I got 3 million of Trumps followers with no problem in a test the other night. To get more than 75k, make sure to set retryonratelimit to TRUE.

answered Jul 25 '17 at 05:00

mkearney

1,266
10
18

Is there a way to fin pagination if the "rtweet_cursor" attr is lost? – CAOC Sep 04 '22 at 18:44

How to get all the Twitter followers ids (> 75000) of a user using pagination in R with rtwitter library?

2 Answers2