Geospatial Analysis of Short-Term Rentals As Determinant of Housing Prices In French Municipalities

Despite the rapid proliferation of short-term rental platforms, quantifying their association with price dynamics remains challenging due to data scarcity. This study employs a spatial data science approach to examine the spatial correlates of housing prices across 25,711 French municipalities. We highlight the distinct advantage of combining official statistics with non-traditional web-scraped sources to create innovative datasets, allowing for the precise characterization of tourism and the uncovering of hidden spatial patterns. Integrating census data with geo-localized 3.5 million Airbnb listings and 1.1 million traditional tourism establishments from TripAdvisor, we utilized exploratory data analysis (Global and Local Moran’s I), unsupervised machine learning clustering (MiniBatchKMeans) and spatial econometric regression modelling (Multiscale Geographically Weighted Regression) to dissect these relationships. The analysis yields two critical insights. First, the relationship between short-term rentals presence and housing prices is spatially non-stationary, with Multiscale Geographically Weighted Regression significantly outperforming global models. Second, while the correlation is generally positive, it is most pronounced in specific high-amenity clusters, such as major urban centres, coastal zones, and Alpine regions, significantly exceeding the association found in rural areas. These findings highlight that the link between the platform economy and real estate values is fundamentally tied to local urban and tourism contexts.