Cleaning up wooly route data

[GavinEllowitz]

I have a layer of several thousand trips data that I would like to use to analyze the popularity of streets in my city's network.  As you can see, the GPS resolution isn't super accurate so I'd like to normalize these as best I can, but I do not know where to start.

[BobBooth1]

It looks like the GPS sample frequency is low, so the segments where two successive points occurred on the same straight street segment follow the network fairly well, but when the street curved, or the driver took a turn, the segments between GPS locations cut across country. 

You might consider extracting all the line vertices to points, and doing a spatial join of the points to your street segments. That way, you could map the number of vertices per street.