Twitter Account Analysis for Drug Involvement Detection

Social media is integral to modern communication, connecting billions of individuals globally. However, this connectivity has also led to negative consequences such as violence, cyber-bullying, cybercrime, and even the promotion of illicit drug-related activities. Among these problems, this paper focuses on an analysis of how drug dealers advertise and drug buyers request drugs using Twitter. For this purpose, we first collect data from users involved in (assumed) illicit drug trade as well as regular users. After monitoring and collecting user data, we store them with our data hierarchy approach and then apply preprocessing, which applies some data cleaning as well as the introduction of additional data from emojis and shared images. This helps better understand and analyze the data. We also apply n-gram analysis to identify patterns and commonalities between sellers and buyers. The results reveal the characteristics of Twitter accounts with (assumed) illicit drug trade including the use of specific words, emojis, and types of photos in their tweets. We observe that Twitter lacks a robust system for consistently monitoring drug-related accounts, with some being suspended while many remain active. Our research and results represent a crucial step toward the development of systems that can effectively identify and address drug-related activities on social media platforms.