Oh well, once you assemble your routes, you will have to determine which intersections are signalled or have stop signs. These data are often available from your municipality. There is little to no point in reducing the street file to an intersection file... only to have 3/4s of the intersections non-signalled. These data would be useful since signal/stop signs often only apply to part of the traffic flow and not flow in all directions. Hence, the need for the data from source or from field work. Or as a simple alternative, you could just get the intersections from the network dataset (it is the points file) and do your multiple select by attributes, summarize as a count and then apply some general penalty for traversal... say 10 seconds per intersection... that doesn't mean 10 will actually be involved at an intersection, but it is a generic penalty. Once you have that, away you go. Sometimes it might just be useful to aportion a penalty to the road distance. This could just be a linear factor. Say travelling along a road with a 60 km/hr speed limit is in reality, like a real speed of 50 km/hr factoring in stops etc. It is up to you. Just do be aware that the actual times you estimate will still have that time component.
So in summary... spending the extra effort may or may not improve your results. What ever procedure you take the results will all be relative to one another and probably bear little resemblance to actual times. That is where actual studies of traffic flow are useful. Good luck
