The problem of evaluating and forecasting the price variation of houses is a traditional one in economic statistics, and the literature dealing with it is very rich. Part of this literature has focused on spatial statistics models in order to account for the structure of spatial dependence among house prices, and studied the relationship between prices and house features, such as dimension, position and type of building. In this paper, we try to extend this approach by considering the effect of exogenous variables, that may exert a significant impact on price dynamics, namely the level of crime and the Airbnb phenomenon. In particular, to our knowledge, the evaluation of the Airbnb activity on the real estate market is still in its infancy, but we expect an increasing role of it. In doing so, we considered the case of New York city, for which this information is fully available as open data, and employed spatial autoregressive and spatial error models, in order to study the impact of these variables along with typical house features on the real estate market for each district of the city.