This is a redistribution of the dataset published on Kaggle by MATTIA ZENI, which is a 30 GB json with youtube video statistics from 2013 sampled using wordnet (it's broad).
This dataset here is parsed into tabular form. The main table is 1millvids_extended.parquet, the primary key is always "X_id" (the unique part of the url).
I ran a linear regression on cumulative views for a sorting that reflects virality (variable "views_cummulative_linear_reg_coef") and added a few key stats on the… See the full description on the dataset page:
https://huggingface.co/datasets/KnutJaegersberg/youtube_stats_one_million.
