Analisis Sentimen Ulasan Google Maps berbasis Natural Language Processing (NLP) sebagai Gambaran Kualitas Pelayanan RS Mata Makassar

Abstract

Introduction: Google Maps reviews provide spontaneous and continuous patient feedback, but their potential for assessing specific dimensions of hospital service quality remains underused. This study aimed to evaluate service quality at RS Mata Makassar through Natural Language Processing (NLP)-based sentiment analysis and the SERVQUAL framework. Methods: This quantitative descriptive study analyzed all eligible Google Maps reviews published from January 2024 to March 2026. Data were collected through manual scraping and processed using cleaning, case folding, normalization, and stopword removal. Sentiment was classified using a rule-based approach into positive, negative, or neutral categories, while service aspects were mapped multilabel to Tangibles, Reliability, Responsiveness, Assurance, and Empathy. Results: The analysis generated 1,495 aspect–sentiment assignments: 1,416 (94.72%) positive, 44 (2.94%) neutral, and 35 (2.34%) negative. Reliability represented the most frequently discussed dimension and was predominantly associated with positive experiences, whereas Empathy demonstrated the highest proportion of positive sentiment. Nevertheless, negative feedback was disproportionately concentrated in the Responsiveness dimension, mainly concerning waiting time, queues, delays, proactive service communication as priority areas for improvement. Conclusion: Google Maps reviews have the potential to complement patient satisfaction evaluations and support continuous monitoring of hospital service quality. Reviews of RS Mata Makassar were predominantly positive, particularly in the Reliability and Empathy dimensions. However, Responsiveness emerged as the main priority for improvement because it had the highest proportion of negative sentiment and accounted for nearly half of all negative aspect–sentiment assignments.