Simplifying (Generalizing) Features and Clustering Points

Question

I'm looking for Katmai functions that will simplify geographic data types.  I saw a demo at some point that showed katmai generalizing a polygon displayed in virtual earth.  It was noted in the demo that this was done during the select - not by modifying the original data.  I'm not able to find what functions may have been used to do this.

Along these lines, how would one go about clustering points with Katmai?  Displaying point data in virtual earth, for example, quickly hits a performance/usability wall at a hundred or so points.  This many points will overlap on the screen, limiting usability, and hindering performance.  What would be the best strategy of limiting the number of results returned based on proximity to one another. 

Thanks for the help.
Erik Schuchmann

Answer 1

You are after the Reduce method. It only works for geometry types and not geography types.

 

As for clustering You could use the STBuffer on each point, find which point has the most points within its buffer and then group those points together.

Answer 2

This has me intrigued. Are you thinking of some custom aggregated method?

If you had to buffer every point, count intersections, determine some logic to choose and then output the recordset I can't see it being very fast. Sounds like nested cursor loops?

 

I work pretty heavily with Virtual Earth and do some manually clustering in .net code currently:

http://www.viawindowslive.com/Articles/VirtualEarth/ClusteringPinsinVirtualEarth6.aspx

 

Would be great to move this to the database.

 

John.