USH BRImo Dataset: Aspect-Based Sentiment Analysis of BRImo Mobile Banking Reviews
Description
The USH BRImo Dataset contains 100,000 Indonesian-language user reviews of the BRImo mobile banking application collected from Google Play and annotated for aspect-based sentiment analysis. The dataset preserves the complete collection prior to data cleaning so that users can independently apply, evaluate, or reproduce different preprocessing strategies according to their research objectives. Each record includes the application identifier, application name, reviewer display name, star rating, review text, posting date, helpful-vote count, developer response, application version, sentiment label, and service-aspect label. The annotations cover three sentiment classes and nine service-aspect categories: customer service, features and innovation, financial value and risk, information quality, other or general experience, security and privacy, system quality, transaction quality, and usability. The repository is accompanied by variable descriptions, annotation guidelines, label definitions, descriptive summaries, visualizations, and baseline aspect-classification results obtained using ten machine-learning algorithms. Data cleaning, validation, preprocessing, and the selection of records for experimental analysis are described separately in the associated data article. The dataset can be reused for Indonesian-language sentiment analysis, aspect classification, joint aspect–sentiment modeling, customer-experience analysis, financial technology service evaluation, and benchmarking of natural language processing methods in a low-resource language context.