engleski

Sentiment analysis and multimodal approach applied to social media content in hospitality industry

Purpose – The importance of the "data gold rush" that occurs in real time on various social media platforms is recognized by various tourism stakeholders and researcher. To extract knowledge from textual data, the purpose of this study is to apply text mining techniques to social media data. Methodology – Descriptive statistical analysis is conducted to quantify the activity of hotel brands on Facebook. The topic modelling technique Latent Dirichlet Allocation (LDA) is used to extract and validate knowledge from text data of 25 Croatian four- and five- star hotel brands that were active on social media in 2019. Sentiment analysis is used to identify personal attitudes expressed through user-generated text that hotel brands promote by posting messages on Facebook pages. Findings – The LDA analysis of the Croatian hotel posts extracted 6 topics: Wellbeing, Atmosphere, Promotion, Gastronomy, Surrounding and Satisfaction. The results of the sentiment analysis indicated that Facebook page followers are more likely to express positive sentiments reflecting an overall satisfaction with the promoted products, services and staff by hotel brands. Contribution – It is a unique study that provides an analysis of textual data in Croatian hospitality research. The application of the multimodal approach contributes to a better uses of contents in possible different strategies so that effective indicators can be given to perform an effective communication. This study provides recommendations, challenges, and current insights into applied communication strategies for marketers to increase a greater number of tourists visiting destinations

tourism ; hospitality ; social media ; text analysis ; sentiment analysis ; Latent Dirichlett Allocation

nije evidentirano