Geoprocessing tool reference > Geocoding toolbox
Understanding geocoding |
|
|
Release 9.2
Last modified August 15, 2007 |
Print all topics in : "Geocoding toolbox" |
In the real world, you find locations based on some description. This might be a number and street name. It might include the name of the city, state, country, or natural features, such as a drainage basin or ecological region.
For example, if you needed to locate the address "380 New York St., Redlands, CA, 92373", with the right street map, it would not take you long to find the exact location. You might first find California, then find the city of Redlands. You might also use a postal code map and locate the region covered by the corresponding ZIP Code value. You would then locate the street and interpret where and on which side of the 300 block the address is located.
Just as you began by narrowing your search to a specific region, finding a particular feature, and interpreting a point along that feature, the computer is doing the same process to assign a location to an address when geocoding. Geocoding starts with a textual description of a location and translates that into the x,y coordinate that can be plotted on a map.
The value of good reference data
Your first step when you want to find something on a map is to have the right map. You will never find your way to 380 New York St. in Redlands, California, if you only have a map of Canada. Also, you will never be able to accurately pinpoint the address if your map only has highways and major cities. Your map needs to have enough detail of the area for you to identify the location for which you are searching.
When you are geocoding in ArcGIS, it is no different. The layers that you use, known as reference data, must also have the details of the specific point you are hoping to find. The primary reference data usually consists of a street network, but a parcel map can be used as well. The important thing is that the data has the detail that you want to find.
An introduction to the address locator
The address locator is the main tool for geocoding in ArcGIS. An address locator is created based on a specific locator style. Once created, an address locator contains the geocoding properties and parameters that are set on the Address Locator Properties dialog box, a snapshot of the address attributes in the reference data, and the queries for performing a geocoding search. The address locator also refers to a geocoding rule base that directs the geocoding engine to perform address standardization and matching.
What the address locator does
Think of the address locator as the sign post covered in signs and arrows, located at the crossroads of geocoding. When you enter an address to be found, the address locator points the input address to a rule base file to be broken into pieces, such as street name, street type, and number; these are known as address elements. The address elements are also standardized, such as making all the directionals or street types the same (for example, Street, St., and Str. would all come back the same). The address locator then sends each standardized address element to a specific category known as a match key. Each match key will then be used for matching and scoring. Some address elements are processed to create index keys, making the search run more easily and quickly. Once possible candidates are identified, individual variable in the candidate is compared with each corresponding address element. A score indicating how well the address is matched is generated. Finally, the address locator presents the best matches based on the score and the location of address being matched. All these paths and directions are determined by properties in the address locator.