Again, please try the more simple example I posted in my previous comment. You find all the code you need (100 lines or so) without dependency to any of my other projection classes. This might help you, at least until I find the time to do some bug-fixing and performance optimization on the net.vis4.renderer package.

One other possibility: could I drop the projection on a sprite, rather than on the stage? The render() method passes a Graphics class as its second argument and I tried creating a Sprite called _canvas, then passing _canvas.graphics to the PolygonRenderer, but it didn't work.

Perhaps I'm not understanding the render() method properly.

I think there might be some bugs in the DataView class.. While doing the map rendering speed comparision I also found out that the map rendering I showed in this tutorial still needs a lot of performance improvements.

So what I recommend to you now is to take a look at a more simple example of map rendering in actionscript. I also uploaded the source file, which contains all calculation needed to render the map. If you want to run this locally, you need to download the used csv map data file as well.

Good luck.

I have a question about the DataView class and the polygon rendering routine, particularly this line:

var screen:Rectangle = new Rectangle(0,0, stage.stageWidth, stage.stageHeight);

When I change the first two coordinates of the screen object to anything besides (0,0), the map no longer draws on screen. Tracing through the code and the DataView class, it appears that the scale (which is incredibly large for my data set) is creating massive x- and y-offsets, so the map is drawn with coordinates in the tens or hundreds of thousands.

Essentially, I want to be able to move the initial starting coordinates of the screen rectangle so my map will draw anywhere on the screen (and at any scale) that I choose. Any suggestions? Was the DataView class only designed with stage-sized boundaries in mind, or am I missing something simple here?

Thanks again.

Let's say you want to project the following coordinates (-180°;-70°), (0°;0°) and (180°;70°) using the Mollweide projection. The projection of the data bounding box would be much larger than the bounding box of the projected coordinates, because at Mollweide the projected longitude distance are getting smaller, the nearer you get to the poles.

No need to calculate the range of the projected coordinates again in coding, because shapefile already has [shp_bounding_box] value stored in for us.

> o = projection.transform(new Point(pt.x, pt.y), new Point());

which is much nicer than using project() because we get results in metres.

Thanks for this remark.

Instead of using
> o = projection.project(deg2rad(pt.x), deg2rad(pt.y), new Point());

Can we use this method
> o = projection.transform(pt, new Point()); in Projection class.