I am having a problem when solving some of my service areas. I am loading the locations from a point feature class, and they are all located. The majority of them return a service area that is appropriate (7 minute breaks, double checked with solving for routes), but a handful of them return something strange. The problem is that there are some areas of the specific service areas that are wildly inaccurate, and there is no way that it is a 7 minute distance, which I also checked. I have attached a picture of the problem, and I'm assuming that it has something to do with my network, but I have no idea what. Any help/ideas would be greatly appreciated.

The network is a walking network that was manually created, and I am using ArcGIS 10.




2 Answers 2


I found that toggling the radio button to "Detailed" instead of "Generalized" fixed this problem.

  • Still puzzled that 'Generalized' would create such results!
    – user173
    Commented Feb 17, 2012 at 11:52
  • 1
    So am I! But I'll accept it for now, as it is still working properly.
    – eric
    Commented Feb 17, 2012 at 21:45

Another option is to play with the "Trim Polygons" setting. This also increases processing time (and more so when combined with a "Detailed" Service Area request) but good results are typically gotten when the Trim Distance is set to 1/4 to 3/4 of your Breaks Distance, or an approximate average equivalent when your Breaks are time and not distance. That is, look at the linear distance your Time Breaks are producing and use 1/4 to 3/4 of that value to set your Trim distance.

Beware that for large projects using the "Detailed" analysis will increase processing time considerably over "Generalized". For small/simple projects it makes little difference.

The reason for spiky-ness when using "Generalized" is that it ignores relatively 'low-flow' routes and, sticking to the 'high-flow' ones, will produce these necks or spikes on the outer edges of the network. "Detailed" makes for a fatter neck thus reducing the spiky character of the resulting outer polygon edges.

Similarly, a sparse network will tend to produce spikes as, from a network model perspective, they look more like a 'high-flow'/"Generalized" network for the purposes of producing Service Area polygons.

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.