Optimizing huge data with kml.

[nige...@gmail.com]

Hi,
I am developing a map application which includes a large amount
of markers to be displayed(Can exceed to 20,000 or even more).
My issue is when I load the map I am also loading a kml file which
includes all the information regarding all those markers.
I fear that if I load such a huge data using kml my site may go slower
or may crash.
Is there any option so that I can optimize my data while loading.
My application is something like a friend circle where I want to
show the locations of all users in the map using icons which on click
popups the information of that user.
Is kml the best practise to load data?
I have read something about clouding markers can anyone help me on
these issues?

Thanks in advance..

[barryhunter]

http://code.google.com/apis/kml/documentation/regions.html

Regions is the KML way organising large quantities of data. Its in
fact used for most of the built in layers in Google Earth.

There are applications around to help you create such a layer, try
searching for 'regionator', superoverlay, or superlayer.

[nige...@gmail.com]

But this will be ideal only in situations where we have different kml
layers to be displayed at different levels.
In my case I am generating kml from postgres table where I am not
sure how to group it into different kml files. So currently I am
generating a single kml file for displaying all my points in map.
What alternative can be done for this?

[vzett]

Regions can also be used to divide the data geographically so that
only a subset of your total points is in view based on the user
viewpoint and zoom in google earth.

In this case, you would determine a reasonable 'sampling' of your data
into lat/lon grids, generate a KML file for each grid cell via
database query, and generate a KML file with region-based network
links.

If you look at my NOAA Wrecks Database layer at http://earthnc.com/awois
- you can see an example of this in operation. Total points is
~20,000 across the entire dataset which is divided into 3x3 degree
grid blocks.

- Virgil
www.earthnc.com
www.maplify.com

[owyheerover]

Another option is to use a network link. The network link can be
configured to send a request back to the server whenever the user
stops moving in Google Earth. The server is programmed to then send a
new batch of data to the network link so that the screen is filled
with data. For an example, try www.earthpoint.us/townships.aspx.
This displays the USA Township and Range grid onto Google Earth.
There are approximately 22 million records in the database. You must
zoom into a western state in order to see the data. As you zoom
closer in or farther out the appropriate degree of detail is
displayed.

Bill Clark
Earth Point

On Apr 15, 6:07 am, vzett wrote:
> Regions can also be used to divide the data geographically so that
> only a subset of your total points is in view based on the user
> viewpoint and zoom in google earth.
>
> In this case, you would determine a reasonable 'sampling' of your data
> into lat/lon grids, generate a KML file for each grid cell via
> database query, and generate a KML file with region-based network
> links.
>

> If you look at my NOAA Wrecks Database layer athttp://earthnc.com/awois

> > > > Thanks in advance..- Hide quoted text -
>
> - Show quoted text -