Figure 2

(A) Distributions of BotScores of Twitter accounts in TWITIMM, i.e., accounts involved in the immigration debate, compared to a random sample of Italian tweets produced during the same period, and equivalent in size; (B) BotScore distribution of all the users in the network, disaggregated by community. We can see that the right tail of the distribution—accounts showing high BotScores—is generally much lower than the left tail. This is particularly true for RT1; (C) Example of BotScore calculated on two different Twitter accounts: one of the authors of the paper (top) and the account with highest U of the TWITIMM dataset (bottom). Results from https://botometer.osome.iu.edu using non-raw scores that span from 0 to 5