Visualizing Point Data

[Bob Basques]

Bill,

The census site I sent you is actually doing this type of overlay right 
now.  It's all MapServer/Mapscript based. Let me know if you have any 
questions.

The overlays are drawn on the fly from the Census BlockGroup data.   
It's got quite a bit of indexing behind the scenes that makes it run 
fairly quickly.   It seems to run out of steam at around 50k points when 
the time seems to take a jump up to 10-15 seconds to process ( the limit 
we were shooting for).  Unfortunately, the code is very Census specific 
at the moment.   We're planning on chopping the code out into a tool 
set, just haven't had time yet.

One other item, the overlays are being drawn right now in a fairly 
course grid, the overlay drawing aspects are configurable and the 
resolution can be increased at the cost of time for processing.

bobb


Bill Binko wrote:

>On Fri, 2006-02-03 at 11:40 -0800, Dylan Beaudette wrote:
>
>  
>
>>I was just working on this last night:
>>http://casoilresource.lawr.ucdavis.edu/drupal/node/172
>>
>>check out the source for v.kernel in GRASS
>>    
>>
>
>Dylan,
>This is an excellent example of what I'm trying to achieve.  Here are
>some quick thoughts about what you've shown:
>
>1) How would you publish this "live" to the web?  Do you know how to
>integrate v.kernel to push its output to mapserver directly?  I know
>there are tools to push GRASS output through GDAL, and I suspect that
>that could be daisy chained into mapserver, but I'd like someone's
>opinion who's actually done this.  If there's an example (anybody?) or
>HOWTO of how to leverage GRASS this way?
>
>2) What is the performance of v.kernel.  For example, I'll have 30-100K
>points over the state of florida.  For a zoomed out view, how long would
>that take to draw?  Perhaps a better question would be, what is the
>order of complexity of the approach taken by GRASS.
>
>3) Unfortunately, because I may want to integrate this into Mapserver, I
>can't quite look at the GRASS source on this.  More specifically, I
>can't copy it since it's GPL into Mapserver, which is BSD and then
>distribute it. Now: I may choose to simply write code to call GRASS
>externally, but I wouldn't need to explore the source for that.
>
>I will however, get GRASS 6 up and running here so that I can explore
>it.  Thanks for the pointer.
>
>Bill
> 
>Bill Binko <bill at binko.net>
>
>  
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.osgeo.org/pipermail/mapserver-users/attachments/20060207/f4cca303/attachment.html