Lab 3 - Determining Quality of Road Networks

In this lab we were to determine the quality of two different networks. For our purposes here, we are determining the completeness of the road networks. We are defining "completeness" for this lab as the length of the road network, so in this lab, the road network that is the longest is the most complete.

First, I wanted to determine the total length of roads for the two networks in kilometers. I added a length field and calculated the total length for each road network. The total length of the Street Centerlines network is 10,805.8 km, and the TIGER network is 11,302.7 km, so overall, the TIGER road network is the most complete. Then, we were to determine the completeness for each grid polygon. This took me some time to determine how to go about this. I initially attempted a spatial join but this does not seem to allow me to compare the lengths per gridcode. I ended up performing an Intersect of both road networks with the Grid layer separately. Then I used the Summarize tool on the GRIDCODE field to get two data tables, which I joined together, and then with the Grid layer, so that I could compare the length of the road networks by their corresponding gridcode. I displayed the data on a map, shown below.

Using the equation found in the lab for the % difference and by using a diverging color ramp, a negative value (shown in blue) means that the TIGER road network is more complete than the Street Centerline network, and a positive value (shown in red) means that the Street Centerlines network is more complete than the TIGER network. The areas coinciding with a more equal level of completeness toward the left center of the map seem to be located near cities such as Medford and Ashland, Oregon (cities not shown on map). Overall, the TIGER network seems to be a little more complete or nearly even with that of the street centerlines network. However, the street centerline network does seem to do a little better near the edges of the map, especially the northwest and western edges. There are also a few strongly positive values in the southeast quadrant of the map.

This was an interesting lab in that it made me think about the differences in road networks and that we take road maps for granted all the time. I think the discrepancies are seen most in the real world when using GPS devices in rural areas or near dirt roads, as they do not see those very well, and I would expect a road network to be most accurate in or near larger population centers.