Source Video Inventory for "Sentiment Analysis and Topic Modeling of YouTube Comments on AI News Anchors Using RoBERTa and BERTopic

This dataset provides an auditable inventory of the 40 publicly accessible YouTube videos used as source material for the study “Sentiment Analysis and Topic Modeling of YouTube Comments on AI News Anchors Using RoBERTa and BERTopic.”The source-video inventory includes YouTube video identifiers, video titles, publication dates, retrieval keywords, candidate-comment counts, and publicly accessible YouTube URLs. The source file from which this inventory was generated contained 7,833 candidate comment records across the 40 videos. Subsequent corpus-construction procedures produced the final analytical corpus reported in the manuscript.This repository provides the source-video inventory rather than redistributing the full text of YouTube user comments. Original comments remain accessible through the corresponding public YouTube video pages, subject to platform availability and policies. User-profile information and author identifiers are not included in this repository.