Covid 19 and lodging places

Abstract

Tourism is a very important source of income for national economies all over the world. Before Covid-19, this sector contributed with 10.4% of the global GDP. Innovative tools for tourism study and promotion are very necessary for a future recovery of the industry. Thus, we have explored Airbnb data as a source of information about the lodging sector, very relevant within the tourism industry. We have analyzed these data to explore the experience of tourists before and after the pandemic. Our aims included identifying and visualizing opinion changes through semantics extracted from semistructured data generated by the Airbnb customers. We used Natural Language Processing and techniques such as sentiment analysis combined with spatial analysis with KDE in order to characterize and spatially visualize user opinion. Results did not show significant differences in user opinion before and after the outbreak of Covid, however spatial patterns related to sentiments were made visible. Moreover, a large dataset covering 3.6M Airbnb lodging spots from 108 cities was compiled and will be made available in the future. This paper can be useful for the lodging industry, tourism organizations as well as social media researchers by providing an alternative approach that involves the role of location in the study of customer behaviour.