Compared to studying other forms of illegal violence, terrorism research faces unique data challenges. Notably, consistent police data are generally lacking, victims often have no direct knowledge of attacks and are sometimes killed, and perpetrators are difficult to interview and are often unreliable sources. In the face of these data challenges, the increasing availability of open source data on terrorist attacks and perpetrators has profoundly affected the scientific study of terrorism. Over time, a growing proportion of published academic research on terrorism—especially more heavily quantitative studies—relies on open source data. Moreover, the growing availability of individual-level terrorism data has allowed empirical tests of psychological and social-psychological theories of radicalization. In this paper, we assess the impact of open source data on the study of terrorism. We begin with an overview, providing a brief history of terrorist attack and perpetrator open source data bases. We then consider the major strengths and weaknesses of both types of open source data and suggest ways of countering the weaknesses. We close with observations about the impact of open source data on terrorism research and future opportunities for improvement.