Interaction Concentration in Public Douyin Posts

抖音公开作品的互动集中与信息竞争边界 Not peer reviewed. This is an executed observational secondary-data analysis, not an intervention experiment. The English and Chinese manuscripts describe the same study. The study analyzes 1,133,545 distinct public Douyin post records associated with 458,271 creators in all 111 partitions of the bendavidsteel/douyin dataset, pinned at revision dd68d28d8cf2605bc3c0f05f62831f9ffc317da8. The highest 1% of sampled posts account for 33.01% of recorded likes, 33.29% of comments, 46.56% of shares and 25.88% of saves. After separately removing the highest 0.1% of observations, the highest 1% of remaining posts still account for 36.08% of shares. The released sample uses keyword-seeded related-post snowball sampling. Its inclusion probabilities and engagement snapshot date are unknown, and approximately 99.9954% of records were created in 2025. Play counts are zero throughout and are not interpreted as measured exposure. These results are conditional on this release; they do not establish that Douyin eliminates information asymmetry, causes economic competition or identifies any named individual's causal responsibility. Data source: https://huggingface.co/datasets/bendavidsteel/douyin . The source is labeled CC BY-NC 4.0 by its releaser. No original post-level dataset, identifiers, profiles or media are redistributed here. The companion ZIP contains acquisition and analysis code, a version-pinned manifest, a locally frozen analysis plan (not publicly preregistered), aggregate results, figure data and numerical validation. The article and original analysis outputs do not change the license of the third-party source data. Author: YANGZHOU. Affiliation: Genbibao Embodied AI Research Institute, Shenzhen, China. ORCID: https://orcid.org/0009-0005-4736-2097 . Correspondence: [email protected] . The author confirms no external research funding and no competing interests beyond the disclosed institute affiliation and marketing business participation. OpenAI Codex assisted source discovery, acquisition and analysis code, execution, checks, figures and drafting. This assistance exceeds copy editing and is disclosed in Methods. AI is not an author.

Authors

Publication Details

Journal
Zenodo (CERN European Organization for Nuclear Research)
Published
2026-11-20
DOI
https://doi.org/10.5281/zenodo.24375252
Primary Topic
Media and Digital Communication
Type
preprint
Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
NOV
preprint

Interaction Concentration in Public Douyin Posts

YANGZHOU
Zenodo (CERN European Organization for Nuclear Research)
Media and Digital Communication
preprint

Interaction Concentration in Public Douyin Posts

YANGZHOU
preprint en

Abstract

抖音公开作品的互动集中与信息竞争边界 Not peer reviewed. This is an executed observational secondary-data analysis, not an intervention experiment. The English and Chinese manuscripts describe the same study. The study analyzes 1,133,545 distinct public Douyin post records associated with 458,271 creators in all 111 partitions of the bendavidsteel/douyin dataset, pinned at revision dd68d28d8cf2605bc3c0f05f62831f9ffc317da8. The highest 1% of sampled posts account for 33.01% of recorded likes, 33.29% of comments, 46.56% of shares and 25.88% of saves. After separately removing the highest 0.1% of observations, the highest 1% of remaining posts still account for 36.08% of shares. The released sample uses keyword-seeded related-post snowball sampling. Its inclusion probabilities and engagement snapshot date are unknown, and approximately 99.9954% of records were created in 2025. Play counts are zero throughout and are not interpreted as measured exposure. These results are conditional on this release; they do not establish that Douyin eliminates information asymmetry, causes economic competition or identifies any named individual's causal responsibility. Data source: https://huggingface.co/datasets/bendavidsteel/douyin . The source is labeled CC BY-NC 4.0 by its releaser. No original post-level dataset, identifiers, profiles or media are redistributed here. The companion ZIP contains acquisition and analysis code, a version-pinned manifest, a locally frozen analysis plan (not publicly preregistered), aggregate results, figure data and numerical validation. The article and original analysis outputs do not change the license of the third-party source data. Author: YANGZHOU. Affiliation: Genbibao Embodied AI Research Institute, Shenzhen, China. ORCID: https://orcid.org/0009-0005-4736-2097 . Correspondence: [email protected] . The author confirms no external research funding and no competing interests beyond the disclosed institute affiliation and marketing business participation. OpenAI Codex assisted source discovery, acquisition and analysis code, execution, checks, figures and drafting. This assistance exceeds copy editing and is disclosed in Methods. AI is not an author.

Zenodo (CERN European Organization for Nuclear Research)
Media and Digital Communication
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.

Laika access coming soon