check with metro on updates

[darrell]

See if there are dramatic changes coming with the next quarterly update

[grant-humphries]

Sent a note to Ben @ Metro about this, so I should hear back in the next day or two

[grant-humphries]

Here's Ben Sainsbury response:

The MAF's issues are mainly due to the volatility/quality of incoming Portland data and then secondarily to the inherent complexity of the dataset.

We're going to include the QA step that we identified a few weeks back (addresses far off the mark) in this next release so, it might not hurt to wait a few weeks, if you can.

I'll let Kellie (Hauger) chime in here also…

A couple other thoughts we had on ways to enhance the import:

1. One thing we do for our locators (not the MAF we distribute) is to identify groups of unit addresses that don't have a 'bare' aka 'parent' aka 'unitless' address counterpart. For each of these identified groups, we then select the first one (aka random), strip off the unit portion and append that to the non-unit addresses. I'm not sure if your code handles this sort of thing, but doing so beforehand, and tossing the unit addresses before the spatial join would probably help improve the output quality.
2. Zac (Christensen) has suggested that as part of the import, you all should consider performing a spatial join of buildings to multifamily polygons in order to pull the mf commonname into OSM when the buildings go in. I realize there may not be a tag for this, but I'm assuming you could create one if needed. The spatial join query should be something along the lines of (PL/SQL)

UPDATE buildings 
SET commonname = 
       ( 
              SELECT commomname 
              FROM   mfhousing 
              WHERE  st_within(buildings.shape, mf.shape)=1)

Definitely looking forward to the final product, good on you for all the hard work,

Don't hesitate to shoot any additional questions our way.

[grant-humphries]

Kellie Hauger of Metro also emailed me with a bit of additional info:

• Metro is within a release or two of adding persisting, unique id's to the address points (the issue of maintenance came up at the last meeting and these could make that process a more viable).
• The city of Portland is in the midst of migrating their address to a new (and improved) database model. This means that there are big improvements coming down the pipe, but it sounds like they are at least a few months away from being released (Kellie recommended that we contact City of Portland for a more finite timeline).
• There will be a new release in the first week of November so we should at least wait until that point. The following release will be the fist week of February

[darrell]

I got this from Neil:

Last week I was in a CGIS (City of Portland Corporate GIS) coordinators meeting. One of the topics was the “Master Address Project” going on at the City and Metro Region. When completed ALL building will be given an “official” address for build and parcel. They will also be handling the one to many relationships (one building to many address – i.e. condos, townhouses, suites….). The intent is to get away from addressing paper documents when assigning address. There will also be a feed-back loop for mailings. Matt Fried is the Manager of CGIS, contact information (Phone: 823-4313 or Email: Matt.Freid@portlandoregon.gov). He should be a able to provide information on the project.

It would seem silly to duplicate the effort. I did bring up using crowd sourcing to check / validate the addressing, but there was no additional information provided.

Also I am under the impression the New LIDAR data flight has been completed and we are waiting for first review of the data. This will have an updated building dataset.

I am looking forward to participating in the effort.

[pdxmele]

Looks like we're all set with the license & etc for now, closing