Analysis of Airbnb Pricing in Lisbon: A Machine Learning Approach: Exploring the Impact of Host Attributes, Property Features and Geographical Factors

Abstract

The socioeconomic impact of Short-Term Rentals has been a hot topic of discussion in recent years across the Portuguese territory. This thesis investigates the determinants of Airbnb pricing in Lisbon using a variety of machine learning models, including Linear Regression, Gradient Boosting and Random Forest. The study aims to understand how different factors such as host attributes, property features and geographical characteristics influence the pricing strategies of Airbnb listings. Data was sourced from the Inside Airbnb website, encompassing a comprehensive range of listings across Lisbon. Key findings indicate that lodging capacity, host response rates, property features such as the number of bathrooms and the type of listing significantly impact prices. Advanced models like Gradient Boosting and Random Forest demonstrated superior performance in capturing complex, non-linear relationships, highlighting the importance of machine learning in predictive modeling. This research provides valuable insights for Airbnb hosts to optimize pricing strategies and for policymakers to understand the impact of Short-Term rentals on the local housing market. Future work should consider incorporating more dynamic data sources, advanced modeling techniques and broader geographical analyses to enhance the understanding of Airbnb pricing dynamics.