I have a built Network Dataset and I have some points that I am creating a service area for. But when I have a short distance to go from the point, it seems to also find routes that is longer than that distance. This mainly happens with two or three of my points because they are not close to the roads. One is around 37 miles from any road, yet it shows that it has a route when I use 30 miles. I've checked everything I can think of, but I'm not sure why this is still happening.
When you load locations into a network analysis layer, the locations have to "locate" on the network dataset. They still display in locations where the original data was (in your case, 37 miles from the nearest road), but the network analysis tools treat them as though they were located directly on the closest street segment. This makes a lot of sense for points that are just a tiny bit offset from the network but much less sense for a case like yours.
When you're in the load locations dialog box, you can change the Search Tolerance, which basically limits how far it will look from your input point to find the nearest road. If you made it smaller, you could tell it not to locate on any roads farther than, say, 30 miles. However, this will simply result in that point being "unlocated," so it will be ignored in the analysis. I'm guessing that's not what you want.
Is your network dataset editable? You might want to digitize in a driveway connecting your point to the road network if you want to get realistic answers.
When you load locations into a network analysis layer, the locations have to "locate" on the network dataset. They still display in locations where the original data was (in your case, 37 miles from the nearest road), but the network analysis tools treat them as though they were located directly on the closest street segment. This makes a lot of sense for points that are just a tiny bit offset from the network but much less sense for a case like yours.
When you're in the load locations dialog box, you can change the Search Tolerance, which basically limits how far it will look from your input point to find the nearest road. If you made it smaller, you could tell it not to locate on any roads farther than, say, 30 miles. However, this will simply result in that point being "unlocated," so it will be ignored in the analysis. I'm guessing that's not what you want.
Is your network dataset editable? You might want to digitize in a driveway connecting your point to the road network if you want to get realistic answers.
