With the emergence of travel and hospitality web sites, travellers can now access and review information concerning their trips and stays before they travel, through user-generated content that undertakes the role of electronic word-of-mouth (eWOM) (El-said, 2020; Le Wang et al., 2020). The main aim in this Thesis is to examine how unstructured information can be combined with structured data structures derived from formal methods for assessing products and services, and, to research possible extensions of these approaches that might lead to more insightful analytics of the Big Data of Tourism, Hospitality and Leisure. The data set for this case study consisted of eWOM posted by travellers in the area of Crete in this case until 2019.

To achieve this, a new multi-dimensional model is developed that includes all the dimensions found in SERVQUAL, HOLSERV and HOLSERV(+) scales. The model achieves to present the information from different hierarchical levels/groupings. Based on the insights gained from the unstructured content corpus’ frequency analysis, additional categorisations wereincluded to improve the model’s efficiency. The final proposed model encompasses online user reviews and structured information derived through mail interviews, descriptive statistics aspect-based sentiment analysis, and multi-class classification, resulting in more sophisticated and insightful data analytics.

Among the findings of this research is the proposed novel Online Review Categorization Model, which is compatible with quality assessment scales, and can be applied in the tourism sector. Moreover, the novel framework developed and applied in this thesis which includes machine learning classification, categorization and annotation approach, with multi-dimensional model development, can be customized and applied to other fields that entail unstructured text that is needed to be classified to multi-dimensional categorization