Logo GISdevelopment.net

GISdevelopment > Proceedings > GITA > 2001


GITA 2002 | GITA 2001 | GITA 2000 | GITA 1999 | GITA 1998 | GITA 1997 |  
Sessions

A tangled web of pure opportunity

Directions for data

Forging the future

How they did it - and what's next

Integrating work management

Mobile solutions- taking it to the streets

Operations support

People make the difference

Systems architecture

The local government perspective

Tying IT all together

Vertical applications


GITA 2001


Direction for Data
Printer Friendly Format

Page 1 of 2
| Next |


Geocoding Technology A Useful Tool For Enterprise Applications

Robert W. Meador
Geographic Data Technology, Inc.
11 Lafayette StreetLebanon, NH 03766


What is geocoding?
Geocoding is the process of locating a street address on a map or assigning geographic coordinates to that address. A "geocoder" - actually a software process - standardizes each street address and assigns it a latitude and longitude based on an Address Coding Guide or ACG. The ACG consists of street network data with address ranges assigned to each block by the U.S. Postal Service. These ranges are frequently assigned in groups of 100 potential addresses per block, and they have parity - that is, odd-numbered addresses are on one side of the street and evens on the other.

Commercial data sets today generally have a "linkage" of around 85% nationwide. This means that the Postal Service can provide address information for 85% of the streets contained in an ACG. Postal information is updated monthly by the Postal Service to add or change street names, address ranges, ZIP Codes and other detailed information. Currency of the USPS information is a key issue in geocoding since most users - utilities included - are very interested in the addresses for new homes and businesses.

When given an address, the geocoder first tries to locate the street on which the address falls. If it does find the street, the software interpolates the address location along the street range. For example, 50 Main Street would be geocoded at the midpoint of the block, and 99 Main Street would be geocoded at the end of the block. If the street can't be found, the geocoder defaults to the "ZIP+4", which is a small, specific delivery subdivision - such as one floor of an office building or one side of a street block - within a five-digit ZIP Code. Addresses that default to a ZIP+4 are geocoded at the midpoint of the block.

Failing to locate the appropriate ZIP+4, the geocoder will attempt to position the address within a group of blocks known as a "ZIP+2" sector. At the largest geographic level, the geocoder will position the address at the "centroid" (or balance point) of the delivery addresses within a 5-digit ZIP Code. For most applications, particularly demographic and marketing analysis, ZIP+2 and ZIP Code locations are close enough. Geocoding software provide statistical information detailing how many addresses were matched at which levels, for example, 75% at the street level, 17% at ZIP+4 level, 5% at ZIP+2, and 3% at ZIP Code level.

Geocoding vendors that offer interactive matching and manual matching services can process addresses that don't geocode automatically. During the interactive phase, map technicians assign additional geocodes by correcting street names and other information from the provided file. For the manual process, digital map technicians call to consult local authorities and use other resources to locate addresses and assign the appropriate codes.

Block-face Accuracy
Some applications require better latitude and longitude information than the interpolated point along a range of possible addresses. "Block-face accuracy" refers the actual range of addresses on a particular block. Using the example above, lets assume Main Street actually has the first house at 45 Main Street and the last house at 75 Main Street. In this case the geocoded point would be interpolated between this range of 45 to 75 instead of the potential range provided by the Postal Service, giving a much more precise latitude/longitude for the actual location. Although not perfect, block-face accuracy eliminates the frequent problem of geocoded addresses bunching up at one end of the street. Obviously, then, it is in the best interest of users to geocode with as much block-face accurate data as possible. GDT is working with a number of utilities and telecommunications companies to improve accuracy based on feedback from the field.

By geocoding your customer information system (CIS) data and displaying it on a map, you can get a quick indication of which streets have "bunches" of geocoded points that are not likely to reflect the real world. Field technicians can then visit the site and determine the correct address range for that street. By providing this address range back to the vendor, you can receive the changed information in the next scheduled update of the data set.

Page 1 of 2
| Next |

Applications | Technology | Policy | History | News | Tenders | Events | Interviews | Career | Companies | Country Pages | Books | Publications | Education | Glossary | Tutorials | Downloads | Site Map | Subscribe | GIS@development Magazine | Updates | Guest Book