Working within a Bayesian parametric framework, we develop a novel approach to studying the distribution of regional population density across space. By exploiting the Gamma distribution, we are able to introduce heterogeneity across space without incurring an a priori definition of territorial units. Our contribution also permits the inclusion of an approximation of individual preferences as a further driving force in location choices. We perform an empirical application to the case of Massachusetts. Our results demonstrate that a subjective measure of distance performs well in replicating the population distribution across Massachusetts.