Accepted for/Published in: JMIR mHealth and uHealth
Date Submitted: Nov 6, 2024
Date Accepted: Nov 19, 2025
Warning: This is an author submission that is not peer-reviewed or edited. Preprints - unless they show as "accepted" - should not be relied on to guide clinical practice or health-related behavior and should not be reported in news media as established information.
“I’m sober and looking to form a community”: Leveraging Natural Language Processing and data mining to explore social engagement dynamics within The Phoenix App, a social networking and recovery platform
ABSTRACT
Background:
Mobile apps are powerful tools for promoting and sustaining healthy behaviors, including those pertaining to substance use abstinence and recovery from substance use. Indeed, prior research strongly supports that social connection via mobile apps, supplemented by in-person interaction, is vital in helping individuals in recovery from substance or alcohol use disorders maintain sobriety as well as for overcoming mental health issues. However, research into the digital footprints of mobile app users, as a strategy to assess app usage experiences in a recovery context, is lacking.
Objective:
This study utilizes the Phoenix mobile app, a social media platform specifically designed for individuals in recovery from substance use disorders, to identify core uses of the app, including how it is leveraged by users from a thematic and emotional valence context.
Methods:
We applied natural language processing and deep learning methods to analyze a random sample of N=19,685 posts. Analyses included the Bidirectional Encoder Representation from Transformers topic modeling tool (BERTopic) to generate themes and a Valence Aware Dictionary and sEntiment Reasoner (VADER) sentiment analysis to approximate emotional tone and mood from posts ranging from highly negative (-.99) to highly positive (.99).
Results:
After removing duplicate and nonsensical posts, we retained a final sample size of N=17,617 posts. BERTopic identified 10 topics (coherence score=.48) within 2 overarching themes: (1) those related to engaging users through in-person and online interactions (7 topics), and (2) as a forum to discuss more serious topics pertaining to substance use and mental health recovery (3 topics). Overall, topics revealed a distinct and recurring theme of community support. VADER sentiment analysis was .44 (SD = .24), suggesting highly positive posts, with only 429 (2.4%) were highly negative.
Conclusions:
Study findings broadly show positive uses of The Phoenix mobile app as a tool for social connection and community among people in recovery from substance use and mental health issues. With the high positive sentiment of posts, the app was distinct from other social media platforms (e.g., X, Reddit, Facebook), which often have a mix of highly positive and highly negative posts. Additional research is needed to confirm these results with a larger dataset and using comparative analysis of other recovery forums in order to contribute to the understanding of social media’s role and function in changing health-related behaviors.
Citation
