Independent Researcher Uploads 4.5B-Record TikTok Video Dataset to Hugging Face After 3-Week Scrape

A three-week scrape of TikTok's Android private HTTP API produced a 4.5 billion-record public dataset with engagement metrics, captions and metadata, uploaded to Hugging Face as 27 zstd-compressed Parquet files totaling ~289GB. The author says the collection used anonymous device registrations with no login required and covers 27 of 32 storage partitions, and acknowledges the pull violated TikTok's terms of service; the dataset card imposes GDPR/CCPA-driven bans on identifying, profiling, or targeting individuals.