I’m building a project requiring large amounts of google maps images. I defined these functions to be used in another function that will automatically collect images. The latitude changes nicely, but I’ve noticed the longitude is slightly off. Is that an artifact of the approximate Mercator projection method? I was under the impression that the conversion I’ve used was pretty accurate except on approaching the poles. The earth is not a true ellipsoid, there are a high number of coordinate systems, and passing from one system to another one is far from simple. You could have a look to pyproj a Python interface to the well known proj.4 library to convert from Lat-Lon (I assume WGS84 …) to almost any other coordinate including of course Mercator. You could try to roll your own, but there are so many caveats such as different origin meridians, slight differences in reference ellipsoid, that you have little hope to have correct and accurate results. I made an object that does similar calculations. Maybe it might give you some inspiration. Basically I treat the earth as an ellipsoid. earthCirc along the equator is not the same as earthCirc through the poles. I try to make conversions between distances in meter <,->, angles of lat & lng. Source.