Machine Learning Analysis of User Sentiments in Tinnitus Management Apps

This dataset contains 342,520 English-language user reviews collected from 84 tinnitus management and related wellness mobile applications available on Apple App Store (iOS) and Google Play Store (Android) platforms, spanning the period from 2015 to 2025.

The dataset was created to support large-scale sentiment analysis and feature-level evaluation of mobile health applications for tinnitus management. Reviews were collected from 15 English-speaking regions using custom Python scrapers, followed by extensive deduplication that removed 5,145,978 duplicate entries from an initial collection of 5,488,498 reviews.

DATASET CONTENTS:- Main CSV file: 342,520 unique user reviews with metadata (author, date, rating, comment text, app name, platform)- Aspect keywords file: Predefined keywords used for aspect-based sentiment analysis across 10 feature categories- README file: Complete documentation of dataset structure, collection methodology, and usage guidelines

The dataset supports research in digital health evaluation, sentiment analysis, natural language processing, and mobile health application assessment. It provides unprecedented scale and temporal coverage for understanding user experiences with tinnitus management apps.

This dataset accompanies the research paper: "Graph Neural Network Analysis of User Sentiment in Tinnitus Management Apps: A Large-Scale Study of 342,520 App Store Reviews" by Yousaf et al. (2026), submitted to BMC Medical Informatics and Decision Making.

COLLECTION PERIOD: 2015-2025GEOGRAPHIC COVERAGE: 15 English-speaking regionsTOTAL REVIEWS: 342,520TOTAL APPS: 84