Ridson Al Farizal Pulungan, Wisnu Adi Agung Nugroho

Aspect-Based Sentiment Analysis on Google Maps Reviews

Introduction

Aspect-based sentiment analysis on google maps reviews. Analyzes West Papua Google Maps reviews with IndoBERT sentiment analysis for tourism aspects. Uncovers post-pandemic recovery gaps, stressing urgent infrastructure & accessibility needs.

Abstract

West Papua is home to several internationally renowned tourist destinations, such as Raja Ampat and Cenderawasih Bay, making the tourism sector one of the key drivers of regional economic growth. However, restrictions on mobility and social interaction during the Covid-19 pandemic led to a significant decline in tourist arrivals in the province. Post-pandemic recovery strategies thus require timely and location-specific data. Google Maps Reviews represent a form of big data that is both up to date and geographically precise, making it useful for assessing and improving the quality of tourism services. This study employs the IndoBERT model for sentiment classification (None, Neutral, Positive, and Negative) across four aspects of tourism: attraction, facilities, accessibility, and price, as reflected in Google Maps reviews. The selected model demonstrates high performance, achieving an F1-score of 71.30% and an accuracy of 93.25%. Findings reveal that the pandemic significantly influenced visitor sentiment, evidenced by a rise in negative reviews during and after the pandemic. This suggests that existing recovery strategies have not been fully effective. Word cloud and thematic map analyses further indicate that the absence or inadequacy of basic facilities and poor accessibility are the primary complaints among tourists. Conversely, the price aspect remained relatively stable, with no substantial increase in negative sentiment, indicating that tourists are more sensitive to service quality than cost. These findings underscore the urgent need for comprehensive improvements in infrastructure and accessibility to support the post-pandemic recovery of tourism in West Papua.

Review

This study offers a highly relevant and timely contribution to understanding the post-pandemic recovery of West Papua's tourism sector. By leveraging Google Maps reviews, the authors address a critical need for location-specific and up-to-date data, providing valuable insights into tourist sentiment across key aspects: attraction, facilities, accessibility, and price. The application of the IndoBERT model for sentiment classification represents a robust methodological choice for processing big data, positioning this research as a practical framework for evidence-based decision-making in tourism development. The research delivers compelling findings, notably demonstrating that the pandemic significantly impacted visitor sentiment, leading to an increase in negative reviews that suggests current recovery strategies are not fully effective. With an F1-score of 71.30% and an accuracy of 93.25%, the IndoBERT model effectively identifies crucial areas of concern. Specifically, the combined analysis through word clouds and thematic maps clearly points to inadequate basic facilities and poor accessibility as the primary sources of tourist complaints. Conversely, the stability of sentiment regarding price indicates that tourists prioritize service quality and infrastructure over cost, offering a nuanced understanding of their sensitivities. The implications of these findings are substantial for West Papua and other regions reliant on tourism. The study provides actionable intelligence, highlighting an urgent need for comprehensive improvements in infrastructure and accessibility to bolster post-pandemic recovery. This research not only showcases a powerful methodology for analyzing online reviews in the tourism context but also furnishes specific, data-driven recommendations that can guide policymakers and tourism operators in strategic planning and resource allocation, ultimately contributing to a more resilient and attractive tourism ecosystem.

Full Text

Comments

You need to be logged in to post a comment.