Service Area Problem - Polygon generation inconsistent

Discussion created by tomcampion on Feb 22, 2012
Latest reply on Jul 16, 2012 by fbelange
Hi everyone!
I am currently working on a walkability analysis identifying various walking distances surrounding schools (200m, 400m, 600m, 800m, 1000m). I have created my own network dataset using multiple datasets and I have not included elevation or direction. This is my first time using the network analyst extension, I have done the tutorials and followed the same process outlined in the tutorials when creating my own network. I am having a problem when solving my service area polygons, most of the polygons that are created look great and are accurate, although some are completely inaccurate and do not represent the appropriate default breaks. Almost as if they cannot traverse properly across the network...??  I have read through most of the threads on this topic already and have not been able to solve my problem. Here is an example of my problem:


This is what I have already ruled out from reading through similar threads:
My facilities are not part of my network dataset.
All of my facilities are located.
There are juctions at every intersection along my network. I also tried dissolve network, still didn't resolve.
I have used the identify tool in the ND extension and all edges reflect the appropriate length.

My polygon generation properties are as follows: Generalised (I have also tried detailed, does not resolve the problem). Trim polygons 50 meters (Have also tried 25 and 100). "Merge by break value" (have tried all others). Overlap type "Rings" (have also tried disks).
Analysis settings as follows. Impedence "Meters" using Shape_Length as my value when creating my ND when selecting evaluators. Default Breaks "200, 400, 600, 800, 1000". Direction "Away from facility". U-Turns "Allowed".

I am lost and am unable to resolve my issue, any assistance to explain why some polygons surrounding facilities are correct and why others are completely inaccurate would be greatly appreciated!