We often take the stance of less is more during turn-by-turn guidance use cases. That means dialing back the map to the essential data points and only contextualizing the ones that are relevant to the current route like turns and lane guidance for example.
I’ve been testing this idea out in the example below. As the car approaches the turn, the camera frames maneuver from a non-pitched perspective for clarity and resolution to the unique geometry at the areas of required action. In this example, we removed all street labels and 3D buildings too, for an ultra-clean and focused driving experience.