Aspect-Based Sentiment Analysis on Generative AI

This dataset contains 1001 app reviews related to generative AI tools, collected from app stores, social media, review sites, and existing publicly available datasets. It is designed for aspect-based sentiment analysis and includes columns for full review text, argument (a specific text span corresponding to a single aspect), identified aspects, sentiment labels, source of the data, and the AI tool being reviewed. A single review may appear in multiple rows when it contains multiple aspects; in such cases, each row represents a unique argument–aspect pair derived from the same review. All personal identifiers have been removed to ensure privacy. Basic preprocessing was performed, including the removal of unnecessary columns and rows and minor text edits. The dataset is intended for research in sentiment analysis, natural language processing, and related tasks.

Version 3 update: This version adds new records to the dataset. A combined file with 1001 total entries (the original data plus the new records) is included. The original file is kept for reference and reproducibility. For any new analyses, users should use the merged file.