I am trying to figure out how to rework my Address Locator or my reference data such that it will return candidates for street segments that have address ranges of 0 - 0 without having to type "0 A ST" for example.
If I type "A St, WROX" (for street name and zone) then I get no results. (This A St does not have any residences or addresses)
However, if I add a "0" in front of the street for an address number, it gives me a 100 match to "A ST, WROX" and strips off the 0 in the match address. Lowering the minimum candidate score does not help.
As a conversation starter: which segment of A st does the geocoder match it to when you add the 0(zero) ?
I have plenty of streets that don't have physical addresses assigned to them, but every segment has a range. (Think freeways: they have a range even if it's based on mile post values and not a single physical address.)
How important is your spatial accuracy/precision when you geocode the phantom addresses, and how many do you have? I wonder if you could just create a set of points that would have A ST, WROX as the matching element and that would work for you. Like a place names sort of feature class. No 'real' address, just a point to match it to.
An alternative is to create a streets feature class of named streets all merged on the street name. Then you could use a US Streets Street Name type of locator. I've never used one because the work I do (9-1-1 dispatch) requires a much higher level of precision with respect to the geocoded result.
Thanks for your reply. Continuing with the A St, WROX example, adding the 0 yields three candidates with 100 score matches. This is because there are three street centerline segments that have the 0 - 0 range, so I believe it is just matching it to the first of the three, not sure how it's choosing which of the three 100 scores to be at the top of the list.
I am preparing for a transition to using Cityworks as a work order management system, so the accuracy of the geocoding is pretty important. There are lots of Public Alleys in Boston that do not have addresses, but there are catch basins and other assets that will have maintenance work orders attached to them, so the yard clerks will need an easy way to geocode and locate the nearby assets quickly. The Public Alley is a similar example because the way the street centerline data and geocoder are set up now, it requires a 0 to be added up front for a match to be found.
I will certainly look into the US Streets Street Name locator type, but it might be helpful to populate all the zero ranges moving forward, to improve accuracy. We have ~172,000 street centerline segments in Boston, and about 4,500 have 0 - 0 address ranges, so less than 3%.