Bayesian modeling of spatio-temporal point patterns in residential property sales

It is of important economic interest to understand the market for sales of residential properties. Customary analysis focuses on explaining selling price using property and neighborhood characteristics, so-called hedonic models. Here, our interest is in understanding the locations of the sales. The set of such locations forms a space-time point pattern. With interest in comparing different types of property sales, we obtain marked space-time point patterns according to the size of property. We focus on sales in the city of Zaragoza in Spain during a 10 year period. We employ nonhomogeneous Poisson process models as well as log Gaussian Cox process models, fitted within a Bayesian ramework, with investigation of model adequacy and model comparison.