Analysis on Customer Satisfaction Dimensions in P2P Accommodation using LDA: A Case Study of Airbnb

Abstract

Customer satisfaction becomes a key influencer for people's habits or daily activities. One of the examples is in the decision-making process about whether they will use specific products or services. People often need other's review or rating about what they are going to use or consume. In this research, by using customer's online review that available from Airbnb website, we try to extract what are the most talked factors about peer-to-peer accommodation, and how customer sentiment about them. We use Latent Dirichlet Allocation (LDA) to extract that factors and conduct sentiment analysis by utilizing semantic analyzer from Google Cloud NLP. We analyze which factors that has more effect on customer satisfaction, not only in general but more specific based on customer gender and tourism destination object. The result shows that factors related to social benefit and service quality have impact on customer satisfaction, moreover different customer gender and different tourism object destination bring different sentiment among customer. We also find several factors that can be improved by the owner of the accommodation to improve customer satisfaction toward their services.

A Geolocation Analytics-Driven Ontology for Short-Term Leases: Inferring Current Sharing Economy Trends

“…The negative effect of short text in the semantic sparseness of the obtained representations may be addressed using clustering techniques. In this setting, short excerpts are spliced into "pseudo"-long texts and subsequently topic-extraction techniques are used in order to identify keywords [30], like latent Dirichlet allocation [31] and latent semantic analysis [32]. A similar objective is achieved through the combination of document-oriented methods with machine learning techniques [33], like bi-directional long short-term memory networks [34], recurrent neural networks and neural language models [35].…”