This paper shows how the implementation of a networkbased approach to predict the location of commercial activities was used in the urban areas of Geneva (Switzerland) and Lyon (France) to support political decision making. Retail stores are qualitatively organized in a network made of links and anti-links showing attraction and repulsion according to commercial categories. The resulting interaction network permits to build a “Q” quality index of optimal location niches for stores, and to classify activities into groups according to their mutual affinities. Maps of both Q-index and groups of commercial activities were produced and used to assess the dynamics and potential effects of an extension of stores’ opening hours in Geneva (tricky political decision). In Lyon, applications of the Q index led to the development of the LoCo software used by the Chamber of Commerce and Industry to advise newcomers on good locations and city mayors on improving commercial opportunities on specific town sectors.