Service Areas definition using Network Analyst

3537
4
10-04-2015 07:43 AM
AlassaneBarro2
New Contributor

I have a road network and a location from which I would like to define a 3-hour service area. I have followed a lab on how to proceed (from time calculation on each road segment, new network dataset creation, loading locations, and setting service area parameters), but I am still getting an unsatisfactory result. However, when I use the lab dataset everything works perfectly. Do you think topological errors in my road network, which I am 100%  sure of the existence, is impacting my results?

Thanks

0 Kudos
4 Replies
ChrisDonohue__GISP
MVP Alum

Yes, I've found with Network Analyst that when things come out not-as-expected, it almost always is an issue caused by either lack of connectiveness or a Network-specific attribute being set wrong.

A quick way to check connectivity is to run another Service Area, but this time turn on the option to "Generate Lines".  Once you have the output, you can visually look at the linework and see gaps.

As for finding and fixing connectivity, run Topology on the Linework.  Be sure to specify rules.

Some Suggested rules:

  • Must Not Overlap
  • Must Not Intersect
  • Must Not Have Dangles
  • Must Not Self-Overlap

Note that you will likely have to do many Exceptions, as there are valid occasions for some of these.  For example, some streets end as a cul-de-sac, so are valid yet will trigger a "Must Not Have Dangles" error.

Topology Rules

ArcGIS Desktop

As for Network-Specific Attributes, review your attributes to be sure that they are appropriate.  Look for things like Oneway set in reverse of expected or a line segment being mis-set by a default attribute as non-traversable/prohibited.

ArcGIS Help 10.1

Give that a try.  If you still run into trouble, post an image of what you got and a description of what you is going awry.  Be sure to specify what version of ArcGIS you are using, as there have been some changes in the last few versions.

Good luck,

Chris Donohue, GISP

AlassaneBarro2
New Contributor

Thanks Chris. I will let you the results once I correct the topological errors.

0 Kudos
JoeBorgione
MVP Emeritus

Chris' comments are spot on. Network analyses demand good data for good results. Clean things up and let us know if it improves your expected results.

That should just about do it....
AlassaneBarro2
New Contributor

Thanks Joe. I will let you know once I clean up the data.

0 Kudos