A compositional analysis of emotions in Airbnb reviews: online and face-to-face experiences

With a focus on personalized travel experiences, platforms like Airbnb have revolutionized tourist interactions by offering authentic local experiences, from cooking classes with locals to city tours. This study addresses the gap in understanding emotions expressed in online versus face-to-face Airb...

Full description

Bibliographic Details
Authors: Marti-Ochoa, Julia, Ferrer Rosell, Berta, Martín Fuentes, Eva
Format: article
Status:Versión aceptada para publicación
Publication Date:2025
Country:España
Institution:Universitat de Lleida (UdL)
Repository:Repositori Obert UdL
OAI Identifier:oai:repositori.udl.cat:10459.1/468062
Online Access:https://doi.org/10.1080/02508281.2025.2489439
https://hdl.handle.net/10459.1/468062
Access Level:Embargoed access
Keyword:UGC
eWOM
Sentiment analysis
DistilRoBERTa
Compositional Data Analysis (CoDa)
Description
Summary:With a focus on personalized travel experiences, platforms like Airbnb have revolutionized tourist interactions by offering authentic local experiences, from cooking classes with locals to city tours. This study addresses the gap in understanding emotions expressed in online versus face-to-face Airbnb experiences. This research utilizes the DistilRoBERTa method and compositional data analysis (CoDa) on peer-to-peer platforms. By examining Airbnb online reviews, this study utilizes sentiment analysis to quantify emotional expressions depending on the type of experiences and categories. The findings suggest that online interactions tend to produce more positive emotions like joy and surprise compared to offline settings, where a variety of emotional responses is evident. Art and Culture, Cooking, and Entertainment categories receive more reviews expressing emotions of surprise and happiness, while Wellness and Tours feature a more varied range of emotions. These methods help us understand the emotional aspects of peer-to-peer experiences, thus improving our knowledge of consumer behaviour.