“Recommend or not Recommend” Topic Modelling of User Reviews in Airline Market

Authors : Fatih Pınarbaşı

Pages : 586-599

Doi:10.25287/ohuiibf.1556680

View : 27 | Download : 24

Publication Date : 2025-04-30

Article Type : Research Paper

Abstract :Understanding customers and the market in the airline industry, which has unique characteristics such as a competitive environment, diverse consumer expectations, and different service levels, is critical for marketing decision-making. Digital platforms offer valuable information sources through online reviews to companies, and it is essential to evaluate the data to evaluate the customers. The study aims to discover the topics in airline user reviews from a duality perspective by recommending reviews and not-recommending reviews. Consistent with the study aim, topic modeling methodology through BERTopic transformers-model is employed to detect the topics included in Skytrax online reviews on Airlinequality.com. 33.810 user reviews from 25 airline companies are used as the study sample. Individual topics detected by topic modeling methodology are grouped into topic groups in the study. Five main topic groups (flight experience, customer service, travel class, airline mentions, and other) for recommending status user reviews, and nine main topics groups (flight experience, service experience, customer service/operations, baggage, customer expressions, region/country-based expressions, seats, transferring process and special cases) for not-recommending status user reviews are concluded in the study.
Keywords : havayolu pazarlaması, kullanıcı önerisi, çevrimiçi değerlendirmeler, kulaktan kulağa pazarlama