Πλοήγηση ανά Συγγραφέα "Kuqja, Franc"

A B C D E F G H I J K L M N O P Q R S T U V W X Y Z

Α Β Γ Δ Ε Ζ Η Θ Ι Κ Λ Μ Ν Ξ Ο Π Ρ Σ Τ Υ Φ Χ Ψ Ω

Τώρα δείχνει 1 - 1 από 1

Airbnb beyond stars: what reviews reveal
(2025-12-18) Kuqja, Franc; Pagratis, Spyros; Dendramis, Yiannis; Dioikitopoulos, Evangelos
Online reviews play a pivotal role in shaping consumer choices on platforms like Airbnb, where textual content offers deeper insights into guest experiences than numerical ratings alone, all while macroeconomic factors like GDP per capita, inflation, and unemployment shape expectations. This study analyzes Airbnb data from eight European cities, first linking pricing to property traits like capacity and ratings, host superhost status, and experience via regressions, then tracing sentiment from review texts over time, spatial city clusters, and economic conditions. NLP converts reviews into sentiment scores, balanced annually per city and aggregated at listing and city-year levels, with sigmoid mapping to Airbnb's 1-5 scale for direct rating comparisons. Token analysis pinpoints words driving positive/negative sentiment, revealing city-specific review tendencies. Key findings: Accommodation capacity emerges as the dominant, consistent price driver, while ratings and host factors show inconsistent impacts. Sentiment uncovers rating inflation—scores cluster at the high end—plus geographic disparities proving uneven guest experiences within cities. Pooled regressions with city fixed effects link economic shifts associatively (not causally) to sentiment fluctuations, though effects remain modest. Ultimately, fusing NLP sentiment analysis, spatial clustering, and macro indicators illuminates platform user dynamics, advancing platform economics and computational social science, despite challenges measuring sentiment in overwhelmingly positive review settings.