Researchers have estimated that 15% of Twitter accounts are bots. While bots are often harmless and bring us joy, there are many bots that are used maliciously, such as spreading false information, which can have negative real-life consequences. Trump, reportedly, has 15 million fake followers that actively contribute to delegitimizing the media by perpetuating his blatantly false claims. Now, more than ever in our history, we must be ruthless in discerning the legitimacy of our information. This chrome extension is one step towards a future where bipartisan beliefs and values can co-exist supported by constructive dialogue based on a single reality.
Thanks to the tremendous efforts of researchers from the Indiana University, there is an amazing body of work around analyzing a myriad of signals (e.g. tweet content and sentiment, network patterns, activity time series) to train highly-accurate models to identify bots.
On top of that, theyʼve even built a wonderful API, Botometer, which accepts a twitter account, its most recent tweets and mentions, and returns a set of scores. Learn more about their API documentation.
Note that bot detection is a difficult task. For example, the current API often categorizes “organizational accounts,ˮ like @BarackObama, as bot accounts. They are also working on improving the API and their training models to better classify bots. To learn more about interpreting the bot score, see Botometerʼs FAQs.
Once youʼve installed the extension, click on the icon so that you can authenticate your Twitter account and provide read-only access. The extension requires it for Twitterʼs API rate limiting purposes.
Then, when youʼre on Twitterʼs website, tweets from accounts that have received a “botˮ high score will be blurred and covered with information about that account.
If you click on the extension icon, there are some simple metrics about the number of accounts scored.
The code base is open sourced here. If you have feature requests or ideas, please feel free to create an issue.
