With the rise of bots accessing social media websites via bot accounts, a need for determining real vs. fake activity is a crucial objective. Botometer, a combined effort of the Network Science Institute (IUNI) and the Center for Complex Networks and Systems Research (CNetS) at Indiana University “checks the activity of a Twitter account and gives it a score based on how likely the account is to be a bot.”
A higher score indicates probability of bot-like activity: 0-to-5 scale with zero being defined as the most human-like and five being classified as the most bot-like. Bayes’ theorem is also used in bot analysis.
A more detailed explanation of Botometer is defined as follows:
“When you check an account, your browser fetches its public profile and hundreds of its public tweets and mentions using the Twitter API.
This data is passed to the Botometer API, which extracts about 1,200 features to characterize the account’s profile, friends, social network structure, temporal activity patterns, language, and sentiment. Finally, the features are used by various machine learning models to compute the bot scores.”
The Botometer service requires Twitter authentication and permissions. This is necessary to check an account’s bot score, Twitter’s REST API is used collect data about a presumed bot account. Subsequently this data is analyzed by Botmeter servers for bot probability examination.
To use Twitter’s API, a user must be have a Twitter account and be authenticated via login. Read/write permissions are used to allow you to block/unfollow users when checking your own followers/friends which is useful to filter and assess Twitter accounts.
Botometer is a useful tool to differentiate between fake (bot originated) and real Twitter accounts allowing users to take appropriate actions regarding the veracity of Twitter opinion and comments.