“Ever searched for a needle in a haystack? It’s a very tedious process if you don’t have a magnet!” That is the metaphor Tom East, GISP, NKAPC’s Senior GIS Specialist, used to describe the process of finding addresses for the multi-purpose address database being built and maintained for use of LINK-GIS partners and emergency dispatch officials.
Although it may seem simple enough to compile a list of known addresses, many problems arise. Sources of the data may record addresses and road names in different ways. For example, MEADOWLARK DR and MEADOW LARK RD may actually refer to the same street, but people may have spelled and abbreviated them differently in their organization. On the other hand, they could also be two entirely different streets.
“It’s our job to resolve these discrepancies and list them as single road names or as separate road names as the case may be,” says East. “The same principle applies to individual addresses along a street.”
GIS staff has used multiple sources in building this address database. Parcel data, utility service address data, and even Google streetview scenes in a few cases, have been used to determine which addresses should be added to the database.
“One of the difficulties is that each source has a large overlap or duplication along with a smaller percentage of unique new addresses that may not be found in other sources. For example, parcel data may list a single address for a property, but utility records could list multiple unique addresses for service on that parcel.”
This fact highlights another issue – that of diminishing returns. Each source must be filtered to eliminate addresses that have been compiled already, while gleaning new ones that might be present. Eventually the number of new addresses discovered by sifting through another source begins to fall flat. Fewer and fewer new addresses can be discovered this way. The most difficult cases may require fieldwork, but as East points out, “We try to use every other affordable means before resorting to fieldwork since it is more costly, both time wise and monetarily.”
Once built, the database becomes a dynamic repository of address information that can be used for multiple purposes. Old addresses are never deleted. Instead they are marked as retired and new ones are always being added.