In all of the 1:10'000 and 1:25'000 examples here, German NAS data was imported into a German-specific AAA data model and automatically generalized. Text was placed automatically as a final step. In these examples, there was no manual intervention at any point. At this scale, building generalization, displacement and data thinning are especially important. Although the data and the model are specific to Germany, data in any model and scale can be automatically generalized.