Now I'm ready to tackle one of my larger problems. I'm working with a cadastral layer and forecasting future development patterns. I have a need to simulate subdivision, where a large vacant parcel (say, 640 acres) will realistically be developed in the future in smaller chunks (say, 1 acre). Right now I ignore the spatial part of it, and end up
I realize that I'm not helping with your specified problem formulation, but that breakup scheme doesn't sound realistic. Besides acreage, there's frontage, and that will typically result in street as well being laid out. I'm sure you realize this and that you're trying to get a rough estimate for some other planning purposes. So, I'd generate a formula that takes a size and estimates the number of lots that can be created, and that might be dividing 80% of the space by the lot size or something, perhaps based on some recent actual subdivision plans with your zoning rules. Then, I'd write some code that takes a large parcel and fully breaks it up into that number of lots (ignoring the street layout issue), and store that infromation, and then construct the residual large parcel based on unioning the lots that aren't yet carved out. The difference in approach is that you'd perhaps divide the parcel into 10x12 chunks rather than taking 1 acre off at a time and getting bizarre residual shapes. Perhaps bizarre residual shapes are appropriate, and you can model the developer holding on to them for 20 years and finally some abutter buying them.... _______________________________________________ postgis-users mailing list [email protected] http://postgis.refractions.net/mailman/listinfo/postgis-users
