Determination of service quality dimensions of online complaints in road passenger transport by text mining

The transport sector is not just about physically getting from one place to another. It represents billions of people, customer experiences, interaction and a huge market. In road transport, intercity bus passenger transport is the largest stakeholder of this market in Turkey. For this reason, customer satisfaction is very important. In order to understand customer satisfaction, customer complaints should be analysed. Today, with the development of social networks, complaints are communicated online instead of classical methods. Social media platforms, online groups and online complaint sites play an important role for customers. Positive and negative comments of customers are very important for companies competing in the market. The aim of the research is to analyse customer complaints in passenger transport, to determine the dimensions of service quality and to identify common problems in transportation. It is to help companies in providing customer satisfaction. The data used in the research were obtained from 7.719 customer complaints made to the three most preferred passenger transport companies in Turkey from the website www.şikayetvar.com using Python programming language. These complaint texts were classified as topic modelling and service quality dimensions using text data analysis methods. Latent Dirichlet Allocation (LDA), an unsupervised classification method, was used to classify the topics. In this study, the data were analysed separately for each company. As a result of the research, it was found that the complaints received by the bus companies from their passengers were caused by physical characteristics, lack of trust on customers and physical characteristics of the buses. Company K received the highest complaint in the dimension of 'Tangibility' with 30.46% compared to other companies. Likewise, Company P, although lower than Company K, analysed the highest complaint in its own data in the 'Tangibility' dimension with 24.4%. On the other hand, Company M received the highest complaint message in the 'Reliability' dimension with 58.93% among the companies and among its own dimensions. Comparisons were made between companies and suggestions were presented in line with customer complaints.