Cannabis Surveillance With Twitter Data: Emerging Topics and Social Bots

Abstract

Objectives. To use publicly accessible data from people who post to Twitter to rapidly capture and describe the public’s recent experiences with cannabis. Methods. We obtained Twitter posts containing cannabis-related terms from May 1, 2018, to December 31, 2018. We used methods to distinguish between posts from social bots and nonbots. We used text classifiers to identify topics in posts (n = 60 861). Results. Prevalent topics of posts included using cannabis with mentions of cannabis initiation, processed cannabis products, and health and medical with posts suggesting that cannabis could help with cancer, sleep, pain, anxiety, depression, trauma, and posttraumatic stress disorder. Polysubstance use was a common topic with mentions of cocaine, heroin, ecstasy, LSD, meth, mushrooms, and Xanax along with cannabis. Social bots regularly made health claims about cannabis. Conclusions. Findings suggest that processed cannabis products, unsubstantiated health claims about cannabis products, and the co-use of cannabis with legal and illicit substances warrant considerations by public health researchers in the future.

Geographic Differences in Cannabis Conversations on Twitter: Infodemiology Study

“…The content analysis done on the US data set revealed some similar themes to previous content analyses such as current use and legalization, transactions, medical use, and the cannabis industry [ 13 , 24 ]; however, our analysis also presented new themes not uncovered in previous content analyses including the topics about the scent of cannabis and places that people use, and medical use and the cannabis industry. Our analysis did not detect several of the themes that other authors did, such as romance, tobacco, and friendship [ 15 ] or processed product use, cannabidiol and hemp use, and polysubstance use [ 24 ]. This could be because previous content analyses were conducted primarily with influential Twitter users, with adolescent users, and not with all available tweets [ 12 , 13 , 15 , 23 ] or because of a more expansive time frame and fewer geographic restrictions than other studies [ 24 ].…”

“…Our analysis did not detect several of the themes that other authors did, such as romance, tobacco, and friendship [ 15 ] or processed product use, cannabidiol and hemp use, and polysubstance use [ 24 ]. This could be because previous content analyses were conducted primarily with influential Twitter users, with adolescent users, and not with all available tweets [ 12 , 13 , 15 , 23 ] or because of a more expansive time frame and fewer geographic restrictions than other studies [ 24 ].…”

“…However, these content analyses had samples restricted to influential Twitter users and adolescents [ 12 , 13 , 15 , 23 ]. Another recent study of tweets from a more general sample captured 12 topics in cannabis-related tweets and also found polysubstance use to be a recurring topic, in addition to topics such as using cannabis, health and medical, cannabis industry, and legality (among others) [ 24 ].…”

Debunking Cannabidiol as a Treatment for COVID-19: Time for the FDA to Adopt a Focused Deterrence Model?

“…In March, the FDA began responding to unsupported therapeutic claims about CBD and novel coronavirus disease 2019 (COVID-19) by sending more strongly worded warning letters that instruct recipients to withdraw such claims within 48 hours [1]. The FDA has limited regulatory capacity to police the numerous CBD retailers and the even larger number of unsupported therapeutic claims about CBD and other cannabis-derived products, so it is worth considering whether this is an efficient, high-impact strategy [2,3].…”

Social scientists battle bots to glean insights from online chatter

“…Since then, Allem has shown that tweets generated by bots are twice as likely as their real counterparts to attest that e-cigarettes help people to give up smoking 1 . Bots are also more likely to tout the unproven health benefits of cannabis 2 . These studies rely on algorithms that estimate the likelihood that a Twitter account is automated.…”