I have list of lat, lng fixes in WGS84, on which I would like to do calculations like distance measurements between point, polygons etc... To do this I was planning on using shapely but then, I need to convert it to a cartesian space, which is only locally accurate.
The problem I have is that my location fixes can come from all over the world, so if I use a fixed projection optimized for my area, I'll introduce errors in other parts of the world. Is it possible to define my own cartesian projection centered around a location pair, depending on the mean position of the current list of locations? The location fixes that I need to do calculations on will always be close to each other, but different lists of location fixes can be spread all over the world.
For example: say I get 5 fixes on which I need to do calculations. Then, I'd like to define a projection which is accurate in the neighbourhood of those fixes, since those fixes wil always be within a few km of each other. When I get the next 5 fixes, which can be on a totally different part of the world, I'd like to define a projection optimized for those locationfixes.
How would I tackle this? It seems like using pyproj (which uses proj.4 if I understood well) is a good idea, but I can't make sense of the string that is required to initialize a projection, as pasted below. Can someone help me?
local_proj = pyproj.Proj(r'+proj=tmerc +lat_0=51.178425 +lon_0=3.561298 +ellps=GRS80 +units=meters')