๐ Digital Ethnography ยท medRxiv Preprint ยท 2026
A PCMCI causal analysis of Tamil community mental health discourse on TikTok โ centering @thadamofficial
Mithun Manivannan, BSc โ Carleton University, Ottawa, Canada
390
TikTok Videos
45
Months
10
Themes
8
Causal Links
โ0.58
Strongest MCI
This research demo is built on top of 29k/Aware โ the free, open-source mental health app by the non-profit 29k Foundation. UI, design system, and components are used under AGPL-3.0. We are deeply grateful to the 29k team for building open, community-first mental health infrastructure.
Visit 29k.org โKey Findings
8 significant causal links discovered via PCMCI (Tigramite, FDR q < 0.05, 500-permutation null model). The headline finding changes how we understand community mental health content strategy.
โ0.58
Self-Harm & Suicide โ Stigma (ฯ=2 months)
Direct crisis content suppresses stigma discourse two months later. Computational proof of the "trust-then-challenge" arc.
+0.33
Self-Harm & Suicide โ Help-Seeking (ฯ=1 month)
Crisis content precedes help-seeking content โ the niederschwelliges (low-threshold) pathway computationally confirmed.
Explore the Research
Explore all 390 videos, the live causal graph, and the 45-month theme timeline โ each directly linked to @thadamofficial's TikTok content.
PCMCI Causal Discovery
8 causal links ยท 10 themes ยท D3 force-directed
Temporal Analysis
45 months ยท Dec 2022 โ Sep 2026
Content Explorer
390 videos ยท search ยท filter ยท TikTok links
Publication
Structured abstract ยท APA ยท BibTeX
SOTA Pipeline (v3)