This thesis discusses problems from the fields of computer vision and congressional districting. The connection between the two seemingly distant subjects is image processing, which can be applied for both skyline extraction and circularity measurement. Hiking applications have a serious problem with the sensor accuracy of mobile devices. With the help of the mountainous skyline and a 3D map, the precision of orientation can be significantly increased. Redistricting has to be carried out to resolve geographic malapportionment caused by the different district population growth rates and migration. This process can be manipulated for an electoral advantage of a party, but achieving optimal partisan districting is not easy at all. In most states of the USA, redistricting is made by non-independent actors and often causes debates about gerrymandering. The highest possible circularity is a natural requirement for a fair legislative district. Thus, shape analysis can be a powerful tool to detect potential manipulation. First, we present an algorithm for skyline extraction and orientation in mountainous terrain, and we also verify the method in a relevant environment. Then, we prove that optimal partisan districting and majority securing districting are NPcomplete problems, and demonstrate why finding optimal districting in real-life is challenging. Finally, we introduce a novel, parameter-free circularity measure that can be used to detect gerrymandering and apply it to congressional districts.