Scraping 3 billion TikTok profiles (and what's actually in them)
The database now holds 3.1 billion unique creators, deduplicated by user id. This is a quick tour of what's inside and how the numbers break down, the sort of thing I'll keep publishing here as the crawl grows.
How many are worth looking at
Most accounts are dormant. If you filter to creators with more than a thousand followers, the count drops to about 344 million, roughly one in nine. That's still a very large addressable set for almost any use case.
Where they are
Country is filled for 99.9% of profiles. The top of the distribution, counting only 1k+ follower accounts:
US 26.8M ID 26.3M VN 13.0M
TH 11.9M BR 11.7M PK 11.0M
PH 11.0M NG 10.8M IQ 8.5MThe long tail is genuinely global, Southeast Asia, West Africa and the Middle East are all heavily represented, not just the US and Europe.
A caveat on country
The country signal comes from the creator's posts, not a fixed "registered region" field. It's usually their home country, but it can drift between posts. If you need it to be right, take it from several recent posts and use the most common value.