Jardim, João Bruno Morais de SousaNeto, Miguel de Castro Simões FerreiraOliveira, Maryana Geovana Bossolani de2025-11-122025-11-122025-10-29http://hdl.handle.net/10362/190603Dissertation presented as the partial requirement for obtaining a Master's degree in Information Management, specialization in Business IntelligenceThe rapid growth of platforms like Airbnb has reshaped urban housing dynamics, especially in tourist destinations such as the Porto Metropolitan Area (PMA). Most existing studies rely on global models that overlook spatial variations in price determinants, limiting the effectiveness of place-based public policies. To address this gap, we applied Multiscale Geographically Weighted Regression (MGWR) to 9,914 Airbnb listings in the PMA, calibrating eight key variables: location and cleanliness ratings, 60-day availability (quadratic term), number of beds², bedrooms² and bathrooms², density of points of interest (POIs) and a tourism accessibility index generated via PCA. The model explained R² = 0.67 of price variation. Cleanliness had a consistently positive effect across the PMA, acting as a universal quality signal. In contrast, availability, bathrooms and amenity density operated at highly local scales, reflecting demand-specific patterns. Structural attributes showed intermediate elasticities, varying with local market competition. We conclude that MGWR is superior in capturing spatial price dynamics. Policy recommendations include: (i) promoting longer stays in central areas to ease housing pressure; and (ii) encouraging tourism diversification in peripheral municipalities. The proposed model offers a practical framework for designing spatially calibrated interventions in short-term rental markets.engAirbnbMGWRShort-term rental pricesPorto Metropolitan AreaSpatial determinantsSDG 8 - Decent work and economic growthSDG 11 - Sustainable cities and communitiesSDG 17 - Partnerships for the goalsmaster thesis204071941