We describe in this communication how we managed to make R, PostGIS and QGis work together to assist us in our econometric modelling of housing prices. We have evaluated different combinations of these three pieces of software that we have used extensively during the course of this research, to finally converge to the system that we are currently using. We were concerned mainly with performance (as large amounts of data are involved), spatial query capabilities, easiness to compute new spatial indicators and to take them into account in the statistical modelling, and finally spatial visualisation of statistical results and map production. As it is often the case with open source software, several possibilities exist to connect the pieces of software two by two. We tried some of them and give an insight on the results of our experiments.