I'm not sure it's entirely useless. By comparing the traffic patterns before and after removing the tollbooth, you could probably quantify pretty accurately the influence it had on the traffic patterns, and then apply that to refine your models of other congestion points.