Precision of geographical coordinates for use in Google Maps
May 22, 2010,
One aspect of using Google Maps that is not often addressed, is the precision of geographical coordinates that is required to properly represent information. A little explanation is in place.
A few days ago a customer mailed me with a problem. She was using the example we have on this blog for a Google map with a marker on it, but she couldn't get the marker to display the correct location. After wasting quite some time fiddling with the marker image and hotspot coordinates, she asked me to look into the problem. It turned out to have nothing to do with the marker itself. The problem was that she specified the geographical coordinates of the location where the marker had to appear in three decimals, not realizing that this simply isn't enough to acurately indicate a location on a street plan.
So, what precision is sufficient? The answer is 'it depends'. It depends on what you're trying to display, the scale of your map, and where in the world it's located. It is more useful to understand how geographical coordinates translate to distance at a local level. If we consider the Earth to be a sphere with a circumference of 40,000km, then the distance between two points with the same longitude, separated by one degree in latitude is about 111km. For longitudes the case is a little different, since the distance between two meridians (lines of constant longitude) decreases as you get further away from the equator. For the high latitudes you may not need as much precision in longitude as you need in latitude, but given that many of the 'interesting' parts of the world are not located at higher latitudes, it is probably best to apply the same precision requirements for longitude and latitude.
Using this information, we can define precision requirements for a few use cases:
- for global maps with large map structures, of the scale of hundreds of kilometers or larger, it is safe to work with one or two decimals for geographical coordinates. Accuracy of the map begins to suffer when the map is zoomed to level 8 or more. The second decimal roughly translates to precision in the order of kilometers.
- for regional maps, at a level of kilometers or more, three or even four decimals are required, allowing reasonably accurate maps to a zoom level of about 13.
- for local maps with accuracy at street level and maximum zoom level, five decimals of precision are required.
Note that at the highest levels, measurement errors start becoming noticeable too. So if the source of your coordinates is something other than Google Maps itself, you might notice slight discrepancies between your coordinates and the map elements. The best way of reducing these is to determine the coordinates in Google Maps itself. Inversely, if you have data of limited precision, it might be sensible to limit the level to whicht your map can be zoomed.
Disclaimer: geographers might take issue with the above, since it makes a few assumptions that aren't exactly true. As a geophysicist by education, I'm fully aware of this, but the estimates made here are certainly sufficient for the purpose at hand.