Appealing and efficient map representations rely on simplifying their map elements in order to reduce the amount of data and rendered objects. However, simplifications inside the well-known OpenStreetMap database are not trivial due to nesting and the associated complex dependencies among elements. A good and consistent simplification of elements in a big way is, therefore, not feasible when working on this data.

This thesis aims to develop a consistent simplification strategy for administrative boundaries inside the OpenStreetMap database. In order to achieve this goal, the relevant map data inside the database gets filtered and transformed into an independent graph representation. Based on this representation, we develop different approaches to simplify, store and unpack the administrative boundaries based on their administrative levels and structure. Therefore, we introduce multiple consistent simplification approaches based on approximation and pruning, a geographical data structure for efficiently storing and accessing geographical data, and different approaches for unpacking simplified data.

Furthermore, we implement a progressive web application that uses the previously defined algorithms and an evaluation proxy for analysing the interaction between components. Finally, all components are benchmarked with static and dynamic analysis regarding the different approaches.