1

As per source code of supercluster, google map utils and mapbox formula to convert (lng, lat) to spherical mercator is as

x = lng/360 + 0.5;
siny = Math.sin(Math.toRadians(lat));
y = 0.5 * Math.log((1 + siny) / (1 - siny)) / -(2 * Math.PI) + .5;

As per the code this is conversion of (lng, lat) to mercator

Result of above formula for spatial point (-122.43499755859375, 37.71860885620117) is {x=0.15990278455946183, y=0.3867186849275602}

If I use PostGIS to convert same point to projection 3857 result is

POINT(-13629401.5834983 4539750.5916923)

Why is there a difference? Leaflet clustering works well with above formula if I have to achieve similar result with PostGIS what kind of projection I must use? Or any more calculations required based on 3857 projection?

  • I assumed (and edited) you meant 3857, not 3875 which is a projection for Finland only – JGH Sep 14 '17 at 12:44
3

The source code you link to seems to be calculating the values in pixels or 0-1 (in one of the comments) as a preparation to multiplying it up to pixels on an image. So the values you have calculated are the proportion of the width/height of the world from the south pole/anti meridian.

To get the Web Mercator coordinates in meters you need to do something like, where C is the circumference of the Earth.

x = C * 0.15990278455946183 - C/2 
y = -C * 0.3867186849275602 + C/2

which for C = 40075000m gives

x = -13629395.9088
y = 4539748.70153

which is pretty close!

  • Web Mercator uses r = 6378137m, so C is 40075016. – AndreJ Sep 14 '17 at 12:46
  • I just used google and took the first answer! – Ian Turton Sep 14 '17 at 12:47
  • @iant thanks, its clear now, but there is one more linked question, bringing in the clustering to PostGIS using web mercator. How do I convert zoom levels to web mercator distance so that I can pass details to ST_ClusterWithIn or ST_ClusterDBSCAN – Srinivas Sep 14 '17 at 13:22

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.