Results from an innovative new study conducted by a team of researchers at Geisel are the first to show that machine learning approaches can be used to identify potential substance use risk behavior, such as alcohol use, among social media users.