Classify users/accounts in Twitter data as bots or not bots.

Returns a numeric vector of probabilities

tweetbotornot(x, fast = FALSE)

botornot(x, fast = FALSE)

botornot_dbl(x, fast = TRUE)

Arguments

x

Object to be classified. Can be user IDs, screen names, or data frames returned by rtweet.

fast

Logical indicating whether to use the fast (lighter) model. The default (fast = FALSE) method uses the most recent 100 tweets posted by users to determine the probability of bot. The fast (fast = TRUE) method only uses users-level data, which is easier to get in large quantities from Twitter's APIS but overall less accurate.

Value

Classifications for all users expressed as probability of whether each account is a bot.

A named (screen names or user IDs depending on input) numeric vector of probabilities

Examples

# NOT RUN {
## vector of screen names
sns <- c("kearneymw", "geoffjentry", "p_barbera",
  "tidyversetweets", "rstatsbot1234", "RStatsStExBot")

## get and view bot probability estimates
twb <- tweetbotornot(sns)
twb

## ask for the fast (user-level data only) version
twbf <- tweetbotornot(sns, fast = TRUE)
twbf

# }