Twitter obviously is a popular platform for communicating and exchanging information. Its popularity, as well as the open API and platform, makes it easy for attackers to exploit and use as a means for efficient distribution of malicious activity. Barracuda Labs has been collecting Twitter data for more than two years and has analyzed more than 20 million user accounts. In this talk, we discuss our findings about the scale and history of malicious activity on Twitter. We measure the Twitter Crime Rate from its inception in 2006 to present day, and then demonstrate how attackers respond rapidly to the large increases of users driven by celebrity attention on Twitter. We also review types of attacks that have been used on Twitter ranging from trending topics poisoning to URL shortener-based attacks. The session then presents work towards building a user-reputation system to statistically identify accounts that resemble fake attacker accounts.