{"doi":"10.1016/j.focus.2025.100408","title":"Content Analysis of Cannabis Discourses on Twitter/X in the U.S.","abstract":"Introduction : With the legalization of both medical and recreational cannabis use in many US states, this study aims to explore public perceptions and discussions about cannabis on social media in the US. Methods : Twitter data on cannabis were collected between February 2022 and February 2023 using the Twitter streaming Application Programming Interface (API). To assess the attitude of tweets toward cannabis and to determine whether Twitter users are cannabis users, human-guided deep-learning models called BERT (Bidirectional Encoder Representations from Transformers) were employed. The sex and age of users were inferred using a deep-learning facial recognition algorithm (DeepFace). The Latent Dirichlet Allocation (LDA) topic model was used to comprehend the topics discussed. Results : Among 2,865,562 unique non-commercial cannabis tweets from the US, 648,018 tweets (22.62%) had a positive attitude towards cannabis, 234,202 (8.17%) with a negative attitude, and 1,983,342 (69.21%) with a neutral attitude. Among 821,451 unique Twitter users, 348,795 (42.46%) were potential cannabis users. US states allowing recreational cannabis use had 12.19 Twitter and cannabis users per 10,000 population, compared to 7.22 in states without it; however, the difference is not statistically significant (P-value = 0.95). The 25-34 age group (37.14%) was the most represented among Twitter and cannabis users. The predominant topic in positive tweets was \"Cannabis's medical value,\" while the main topic in negative tweets was \"Having difficulty quitting cannabis.\" Conclusions : Our study provides a detailed overview of public perceptions of cannabis in the US, aiding policymakers and public health authorities in developing effective regulatory policies about cannabis.","journal":"AJPM Focus","year":2025,"id":545856,"datarank":0.0,"base_score":0.0,"endowment":0.0,"self_citation_contribution":0.0,"citation_network_contribution":0.0,"self_endowment_contribution":0.0,"citer_contribution":0.0,"corpus_percentile":null,"corpus_rank":null,"citation_count":1,"citer_count":0,"citers_with_citation_signal":0,"citers_with_endowment":0,"datacite_reuse_total":0,"is_dataset":false,"is_dataset_confidence":0.7687,"is_data_producer":false,"deposit_databanks":null,"is_oa":true,"file_count":0,"downloads":0,"has_version_chain":false,"published_date":"2025-01-01","fair_score":null,"fair_percentile":null,"algorithm_id":"datarank_citation_only_1hop_v6","ranking_scope":"data_only","authors":[{"id":940867,"name":"Runtao Zhou","orcid":"0000-0002-1745-2148","position":1,"is_corresponding":false},{"id":1437564,"name":"Qihao Yun","orcid":null,"position":2,"is_corresponding":false},{"id":1437565,"name":"Jingzhu Wu","orcid":null,"position":3,"is_corresponding":false},{"id":610983,"name":"Zhengyuan Wang","orcid":"0000-0003-2724-6396","position":4,"is_corresponding":false},{"id":1437090,"name":"Minggang Yu","orcid":"0000-0002-2732-3219","position":5,"is_corresponding":false},{"id":225749,"name":"Karen M. Wilson","orcid":"0000-0002-2703-7001","position":6,"is_corresponding":false},{"id":278344,"name":"Dongmei Li","orcid":"0000-0001-9140-2483","position":7,"is_corresponding":false},{"id":305192,"name":"Zidian Xie","orcid":"0000-0002-5149-7710","position":0,"is_corresponding":true}],"reference_count":44,"raw_metadata":null,"created_at":"2026-07-19T02:53:27.751050Z","pmid":"41035944","pmcid":null,"fwci":null,"citation_percentile":null,"influential_citations":0,"oa_status":null,"license":null,"views":0,"total_file_size_bytes":0,"version_count":0,"fair_f":null,"fair_a":null,"fair_i":null,"fair_r":null,"fair_zscore":null,"fair_rationale":null,"fair_model":null,"fair_agent_version":null,"fair_fulltext_source":null,"fair_has_llm":null,"fair_computed_at":null,"clinical_trials":[],"software_tools":[],"db_accessions":[],"linked_datasets":[],"topics":[]}