Siap Kerja Indonesian Google Play Review Sentiment Dataset: Public No-Text Reproducibility Package

This record provides a public no-text reproducibility package for a study on multiclass sentiment classification of Indonesian Google Play reviews of the Siap Kerja application, a public employment-service platform in Indonesia. The package supports verification and reconstruction of the reported experimental results while reducing privacy and platform-related risks associated with redistributing user-generated review text. The package includes anonymized internal review keys, no-text metadata, final sentiment labels, split manifests, final-test prediction files, evaluation outputs, confusion matrices, bootstrap comparison results, data-integrity summaries, and documentation needed to verify the reported model performance. The sentiment labels consist of negative, neutral, and positive categories. Raw review texts, usernames, profile images, original review identifiers, and individual text-bearing annotator files are not included in this public package. The package is intended to support transparent, privacy-aware, and reproducible evaluation of classical machine learning and transformer-based models for Indonesian public-service app-review sentiment classification.