mpadge
commented
6 years ago This is a great idea, and I agree likely to have useful benefits. I've also not really tried it, but it should pretty much work as is. An osmplotr map is just a standard ggplot2 object, so any additional geoms/stats should just work on top of it. The only thing osmplotr does differently is the final print_osm_map() function, but that shouldn't affect any of this. We'd just need some good examples to get going.
I've been thinking about testing some special geoms/stats for osm objects. There may be some useful benefits from doing things that way in terms of flexibility, but I've only written a couple previously so I'm not sure of the implications for how the rest of osmplotr works. e.g. might not be compatible with the mechanisms for supplying defaults.