Help with design patterns for geodata centric app.

[Ryan Goodrich]

Hello, I am trying to figure out the best way to design the internal working of an app that is heavily based on alot of geocoded data. I have run into and area that is somewhat illusive to me and would really appreciate anyones wisdom or guidance on how best to tackle this.

What I am trying to figure out is the issue of zooming out. My app will have thousands of points all over a map represented by pins. This poses the issue that when you zoom out, your going to eventually end up with a thousand pins on screen, and one seriously huge geoquery return. So some system has to be created to deal with this by ideally clumping together of geopoints that are nearby. I've been investigating the stack mob docs and cannot find any built in technique for dealing with this, if there is one please point me to it's direction, but I couldn't find it. So I am assuming I am going to have to get clever and do something from scratch. Which is where it starts to get illusive to me, I have about some ideas on how to accomplish this, but am not sure which would be best, or if theres just some method that I am missing entirely. 

The method I am starting to rest on is one where, I will split the viewable area of the map up into lets say 64 equally spaced rectangles (8x8). I will then fire off 64 queries and have it only return 1 result per query to see if something is in the squares. If something is in there, it will display 1 pin. Once the user zooms in enough, it will do this again, and again. until finally when the user has zoomed in enough, it will do away with this method and just geoquery all the points and show them as they are. 

Is this a good idea? Or is there a better way to do it?