A Comparative Automated Text Analysis of Airbnb Reviews in Hong Kong and Singapore Using Latent Dirichlet Allocation

Abstract

Airbnb has emerged as a platform where unique accommodation options can be found. Due to the uniqueness of each accommodation unit and host combination, each listing offers a one-of-a-kind experience. As consumers increasingly rely on text reviews of other customers, managers are also increasingly gaining insight from customer reviews. Thus, this present study aimed to extract those insights from reviews using latent Dirichlet allocation, an unsupervised type of topic modeling that extracts latent discussion topics from text data. Findings of Hong Kong’s 185,695 and Singapore’s 93,571 Airbnb reviews, two long-term rival destinations, were compared. Hong Kong produced 12 total topics that can be categorized into four distinct groups whereas Singapore’s optimal number of topics was only five. Topics produced from both destinations covered the same range of attributes, but Hong Kong’s 12 topics provide a greater degree of precision to formulate managerial recommendations. While many topics are similar to established hotel attributes, topics related to the host and listing management are unique to the Airbnb experience. The findings also revealed keywords used when evaluating the experience that provide more insight beyond typical numeric ratings.

Sentiment Analysis and Multimodal Approach Applied to Social Media Content in Hospitality Industry

“…To perform LDA and define the optimal number of topics, the authors evaluated the performance of different selections of K as the total number of topics. As suggested by Kiatkawsin et al (2020), the results of each method were normalised on a scale from 0 to 1 before the LDA models ran the results. It is worth noting that every possible number of topics between 2 and 50 was identified and scored through a grid search.…”

“…The results of LDA showed that there are five aspects of the hotel: location, meal, service, comfort and cleanness, and concluded that the overall sentiment is influenced by one aspect. Kiatkawsin et al (2020) compared two rival destinations, Hong Kong and Singapore, using latent Dirichlet allocation. In their research, 185,695 from Hong Kong`s and 93,571 from Singapore`s Airbnb were analysed, and a total of twelve topics were identified for Hong Kong, which are be categorized into four different groups, while Singapore's optimal number of themes was only five.…”

“…However, since LDA does not explicitly state the topic names, authors must interpret the overarching topics of the keywords in each topic and specify names. The validation of the topics was accomplished by analyzing the Facebook posts of the hotel brands with the highest proportion of extracted keywords in the relevant topic to confirm that the topic labels appropriately reflect the topic content (Kiatkawsin et al 2020). The metrics od "Deveaud2014" and "CaoJuan2009" indicate that the optimal number of topics in the sample of Croatian hotel brand posts is between 2 and 6 topics.…”

The Moderating Roles of Sensation Seeking and Worry among Nature-Based Adventure Tourists

“…Previous experience and existing knowledge related to adventure activities were found to have significant influences on attitude [ 6 , 23 , 26 , 27 ]. Thus, attitude helps to determine the likelihood of participation in the activity [ 28 , 29 , 30 , 31 , 32 ]. Moreover, constructs such as sensation seeking and worry have been found to help in coping with the level of perceived risks usually associated with adventure tours [ 33 , 34 , 35 , 36 , 37 ].…”

Classification and Geotemporal Analysis of Quality-of-Life Issues in Tenant Reviews

“…Text classification refers to the process of categorizing textual data into a set of defined classes. Classical approaches to text classification rely on feature extraction techniques such as n-grams, Bagof-Words, and TF-IDF, a potential dimensionality reduction step, followed by learning a classification model such as Logistic Regression, Naive Bayes, Support Vector Machines, Latent Dirichlet Allocation, and Nearest-Neighbours algorithms (Kowsari et al, 2019;Kiatkawsin et al, 2020). More recently, deep-learning-based language models that are trained using contextualized word representations have been used to achieve state-of-the-art results on a wide range of natural language benchmarks and datasets, including text classification (Devlin et al, 2019;Lewis et al, 2020;Minaee et al, 2021;.…”