Setting up travel direction when creating network dataset?

Question

I want to conduct service area analysis with a roadway network to create travel time contour map. The roadway netwotk is a linear shapefile which stores travel direction (N/S/W/E) and travel time for each roadway segments. So for a two-way road, there will be two lines with opposite travel direction and different travel time. So different lines would be used in calculating travel time from and towards the facility.

However, I don't know how to setup travel direction when creating the network dataset by using the existing travel direction.

Maybe I can set the travel direction of all lines as one-way and let GIS to determine the travel direction? However, the digitized direction of the lines are not the same with the travel direction.

Answer 1

Cardinal directions aren't really considered when talking about travel direction on a network edge. See my answers here and here. The situation you're describing is more like traffic at a given time than one-way restrictions. It would be possible to model this by having a separate to-from and from-to cost for the same edge. This would be set up in the drive-time evaluator under the network properties attribute, as I described in my answer to your previous question. No one-way or creating two edges for the same road would be necessary, it can all be done with a single edge and having two attribute fields - one for travel time one way and one for the other direction.

I suggest taking a look at the Network Analyst Tutorial in the help files, particularly Exercises 1 and 12.

• 12B01: Thanks, Chris. The problem is that the speed data I get is stored in the roadway network with separate lines represents different travel direction. It is also for the mapping purpose (not travel time contour map, but map shows detail about speed in each direction). Am I right that if there is no any travel restriction, the results of from and towards facility generated by service area analysis would be the same?
• Chris W: @12B01 That would depend on the network geometry and difference in direction speed, but yes, in theory it'd be the same. The analysis is going to take the least impedance route and without a one-way restriction it could possibly switch between the two sides unless you're talking about stacked, identical lines. In that case, it will for sure use the lesser values regardless of start point without a restriction. And you'll have to get travel relative to dig direction as that's how the software functions. It sounds like there might be a lot of cleanup required to use that data.
• 12B01: Yes, I think the dataset I get is not ideal for service area analysis. There are lot of preparatory work to do. Thanks for your help Chris!