Abstract (Translated):
This study makes a contribution to the real estate market in Argentina by utilizing online information and applying statistical models based on the hedonic approach. Using data from over 63,000 apartment listings for sale, three statistical methods are applied: Ordinary Least Squares (OLS), Least Absolute Shrinkage and Selection Operator (LASSO), and Classification and Regression Trees (CART). The explanatory variables used are grouped into four categories: property characteristics, neighborhood attributes, geographic proximity, and mobility.The results reveal that attributes such as size, number of rooms, and the presence of amenities have a positive and significant effect on prices. Location in high-value neighborhoods, such as Puerto Madero, is also a determining factor. Proximity to green spaces and universities is associated with higher prices, while closeness to public transportation has a negative impact. In terms of mobility, greater circulation in residential areas is linked to higher prices, whereas mobility in commercial and entertainment areas has the opposite effect. Among the methods used, LASSO demonstrated the best predictive performance for out-of-sample observations.These results should be considered not only from a quantitative perspective but also from a social standpoint, given the relevance of housing as a key asset. This information is useful for decision-making in housing policy.
