[Imports] [Talk-us-newyork] [Imports-us] Update: New York State GIS SAM Address Points Import

[Skyler Hawthorne]

Hi Jack,

Thanks for reviewing the data! This appears to be a bug due to a case I did not think about. I recently discovered that some of the data has a Floor field that indicates which floor of a building an address is on, and added this field to addr:floor when parsing (which reminds me I need to update the wiki page). In other cases I've seen with Floor, they were on addresses on apartment buildings, and they also had units. This makes sense: it is telling you which floor the unit is on, and it adds useful information. So I wrote the algorithm that combines them under the assumption that if there's a point with a floor, it's accompanied by a unit. This yielded groups of units that indicate which floor they are on, such as these two on this building:

https://www.openstreetmap.org/node/8394580064

https://www.openstreetmap.org/node/8394580063

However, it appears these buildings have separate address points where there are separate points that have units and floors without the other. It isn't terribly useful to have an address point with just a floor without a unit, because now we can't tell which unit is on which floor without a survey. So I suppose we could either keep these and mark them for review, so that someone could do the survey and deduplicate, or we could just drop these address points that only have a floor and no unit.

I think I'm leaning toward the latter.

Thanks for helping find this bug!

--
Skyler

Feb 6, 2021 16:03:48 Jack Arnold via Imports <imports at openstreetmap.org>:

> Hi Skyler,
> 
> Looking over some of the address points added in Albany and they look
> good! I'm not from the area, but I noticed some weird behavior for what
> I believe is a large apartment/flat:
> 
> Building footprint way:
> https://www.openstreetmap.org/way/865748654
> 
> Nodes from the import:
> https://www.openstreetmap.org/node/8399603901
> https://www.openstreetmap.org/node/8399616671
> 
> The building footprint way has a different address, so the address
> point was added over it. There appears to be two points added over it
> in the same spot, each with combined nysgissam:nysaddresspointid that
> are different. If the total nysgissam:nysaddresspointid length was less
> than the limit, why didn't these nodes get combined?
> 
> Again, I don't know if this is the correct behavior, just something I
> noticed that appears across this whole block.
> 
> Thanks for the good work on importing address points, I look forward to
> being able to use them!
> Jack