Most of our knowledge of the factors favoring pollinating insects in urban areas arises from studies focused on urbanization gradients. However, when it comes to differentiating among types of built fabric according to their capacity to host pollinators, few information can be found. In this paper, we study the influence of built fabric patterns on pollinating insects using open data from citizen science and covering the whole of France. With pollinator richness as the response variable in a model focused on urban agglomerations, we test the relevance of spatial metrics describing the building patterns while accounting for other drivers such as plant species, season, climate, and the connectivity of herbaceous habitats. We divide the data into subsets covering varying ranges of building densities, to explore those for which built fabric types exhibit the greatest differences. Our results show that only a focus on moderate building densities reveals clear differences between built fabric types. In this case, the areas with old individual housing appear to be more favorable than other patterns dominated by individual housing built in the post-war period and in the late twentieth century. These results outline the potential of the coupling of participative naturalist data with geographical data.