An alternative approach to understanding airline customers’ attitudes and behaviors

The aviation industry is facing challenges in meeting the evolved expectations of travelers in the post-COVID era in which their expectations are expressed in their online reviews. Unprecedented proliferation of passenger reviews on online platforms has resulted in a deluge of unstructured data. Hence, in order to assess customers’ attitudes and behaviors, this study utilized machine learning technique including topic modelling and sentiment analysis through analyzing passenger review data extracted from Skytrax for the top 100 low-cost and flagship airlines in 2024. Two most popular topic modelling methods, Latent Dirichlet Allocation and BERTopic were applied to identify recurrent themes. The results show that flight delays, schedule-related issues, communication gaps, and lack of pricing transparency were underscored by passengers as major shortcomings. Low-cost carrier passengers were more focused on operational reliability; whereas, flagship airline passengers were mainly concerned about extra charges and quality of services. The sentiment analysis reveals that cultural and regional differences have influenced customer satisfaction regarding the airline types, continents, cabin comfort, handling of baggage, and in-flight service provision. This research contributes to existing service quality research by extending the traditional methods of assessing customer attitudes and behaviors. The findings provide insightful guidance for airlines to improve service offerings and attract customer patronage.