Geocoding & Route Mapping

This week I was given a set of addresses and was required to geocode them, i.e. associate them with their proper locations on a map. In order to do so, I downloaded the TIGER/Line file from the U.S. Census Bureau for Lake County, FL, which is where the addresses were located. The TIGER/Line file is a topologically correct vector line dataset for streets and addresses in the U.S., which means that the dataset is made up of edges (usually road segments) and junctions (crossroads) such that a path or paths can be calculated from one junction to another. The TIGER/Line file also associates house numbers with street segments, as well as zip codes and a wealth of other addressing data.

To geocode my set of addresses, I ran a geocoding tool in ArcGIS, which correctly placed 68% of the addresses. A 32% miss rate is not acceptable, so I had to look at the addresses that went unmatched and manually locate them (with the help of google maps in this case). I was able to do this successfully with all but one address that seemed to have been misspelled, leaving me with a 95% success rate, which was good enough for the purposes of this exercise.

With my geocoded data, I was then able to use ArcGIS' Network Analyst tool to do some simple routing analysis, creating a route based on three arbitrarily selected road stops. A route was then created that minimized travel time and hit all three stops.

Optimal Routing Example for Lake County, FL