A Comparative Text-Mining and Semantic Network Analysis of User Reviews on Traveloka and Trip.com

Abstract

The global online travel market surpassed US$820 billion in 2024 and is expected to reach US$1 trillion by 2027, fueled by digitalization and the rise of mobile-based bookings. Within this expanding sector, Trip.com and Traveloka stand out as leading online travel agencies (OTAs) that depend heavily on user-generated reviews for service enhancement and competitive positioning. This study employs a mixed-method approach combining text mining, and semantic network (CONCOR) analysis to examine 2,000 user reviews (1,000 per platform) gathered from both OTAs. Using RStudio, UCINET, and CONCOR, the analysis uncovers key linguistic and thematic patterns that reflect users’ perceptions of service quality, usability, and issue resolution. Findings reveal four shared thematic dimensions: Platform & Travel Logistics, Complaint & Problem Handling, Travel Experience & Services, and Transaction & Booking Process. However, emphasis differs: Traveloka reviews stress refund processing, reliability, and responsiveness, while Trip.com reviews highlight booking efficiency and trust issues, particularly regarding refunds and cancellations. This research advances the application of big-data text analytics in comparative OTA evaluation and demonstrates how semantic relationships in user reviews uncover service perceptions. Practical implications include improving transparency, refund management, and leveraging AI-driven personalization to enhance customer satisfaction.