Related Experiment Video
Updated: Sep 9, 2025

Integrating Computerized Linguistic and Social Network Analyses to Capture Addiction Recovery Capital in an Online Community
Published on: May 31, 2019
Decoding HIV Discourse on Social Media: Large-Scale Analysis of 191,972 Tweets Using Machine Learning, Topic
Xiangming Zhan1, Meijia Song2, Cho Hee Shrader3
1Edson College of Nursing and Health Innovation, Arizona State University, 500 N 3rd St, Phoenix, Phoenix, AZ, 85004, United States, 1 (330) 272-4294.
Social media, particularly X, is a key source of HIV information due to barriers to healthcare access. This study analyzed 191,972 tweets, revealing dominant informational themes and distinct engagement patterns for awareness campaigns versus community events.
Area of Science:
- Public Health
- Digital Health Communication
- Social Media Analysis
Background:
- Social media platforms are crucial for HIV information dissemination due to stigma and healthcare access barriers.
- Existing research on HIV discourse on social media is limited in scope, scale, and timeliness, especially post-COVID-19 and Twitter's rebranding to X.
- Understanding the comprehensive HIV-related discourse on X is vital for effective public health strategies.
Purpose of the Study:
- To identify dominant themes in current HIV-related social media discourse on X.
- To compare theory-driven and data-driven thematic analysis approaches for HIV content.
- To examine the influence of emotional responses and temporal patterns on HIV content dissemination.
Main Methods:
- Analysis of 191,972 tweets from June 2023 to August 2024 using an integrated framework.
- Utilized supervised machine learning for text classification and comparative topic modeling (theory-driven vs. data-driven LDA).
- Employed sentiment analysis (VADER, NRC Emotion Lexicon) and temporal trend analysis to assess engagement.
Main Results:
- Information and education content comprised the majority (63.02%) of HIV discourse, followed by opinions (12.43%) and personal experiences (10.25%).
- Data-driven topic modeling identified 8 distinct themes, with some overlap and unique findings compared to theory-driven categories.
- Official campaigns (e.g., World AIDS Day) showed delayed engagement, while community events (e.g., National HIV Testing Day) prompted immediate peer-to-peer interaction.
Conclusions:
- HIV discourse on X is dominated by informational content, with prevention emerging as a key focus.
- Effective HIV communication requires integrating medical information with community perspectives and strategic message timing.
- Findings offer guidance for digital outreach strategies to connect vulnerable populations with healthcare services in the post-pandemic era.
Related Concept Videos
Steps in Outbreak Investigation
Statistical Methods for Analyzing Epidemiological Data

