datasocial/tiktok-5.6B-videos
massive TikTok video dataset
1 min read
This large-scale dataset features information on 5.6 billion TikTok videos, including detailed engagement statistics, post metadata, hashtags, and audio tags. Built for research purposes and personal use, it lets you analyze trends and user behavior patterns. It is an ideal fit for data scientists and researchers studying recommendation systems, social media trends, or NLP tasks based on user-generated content.
Key features
- Massive scale: data on 5.6 billion TikTok videos and photo posts.
- Detailed columns: video ID, post date, country, language, duration, resolution, and content type.
- Real-time stats: views, likes, comments, shares, saves, and a pre-calculated engagement rate (engagement_rate).
- Extra markers: tags, creator mentions, on-screen text, ad flags, and TikTok Shop integrations.
What you can do
- Explore trends and recommendation algorithms of popular video platforms.
- Train analytical models and classifiers for working with social media text and metadata.
- Conduct marketing analysis of user behavior and viral content.
Example
Loading the dataset using the datasets library in Python:
from datasets import load_dataset
dataset = load_dataset("datasocial/tiktok-5.6B-videos")
print(dataset["train"][0])
Read more
DetailsScale: 5.6B records, including archived videos.
Format: tabular data (tabular, text) with numerous columns (video_id, posted_at, country, engagement_rate, etc.).
License: CC BY-NC 4.0 (free for research and personal use).








Comments