A Case Study in Twitter Bot Identification: Are They Still a Problem?

Tiange Wang, Fengkai Wu, Richard O. Sinnott

Abstract: Social media is used by billions of people and organisations worldwide for entertainment, advertising, news and many other reasons. It is also often used for gauging public opinion and spreading of information. This information can directly influence public opinion. Unfortunately, the potential for automatically influencing public opinion by potentially nefarious organisations and indeed countries through flooding of posts and tweets is increasingly common and of concern. Bots are often used to automate the sending of this information and potential misinformation. In this paper we explore approaches for bot identification in an automated manner using extensive Twitter collections related to the city of Melbourne. Despite the work of social media platforms such as Twitter in cleaning up the use of their platform and minimizing the widespread use/misuse of bots, we identify that a very small percentage of bots may still be active and must be addressed.