[Talk-us-massachusetts] Braintree buildings+

[Greg Troxel]

"Wayne Emerson, Jr. via Talk-us-massachusetts"
<talk-us-massachusetts at openstreetmap.org> writes:

> MassGIS announced on August 28, 2019 an update to 2D building
> structures, describing it as a, "Comprehensive statewide refresh to a
> baseline of 2016, plus data for Boston from the city."
> https://docs.digital.mass.gov/dataset/massgis-data-building-structures-2-d

Thanks for pointing that out.

> I find a good way to compare our OSM buildings with MassGIS' is to
> open the Massachusetts Interactive Property Map at
> http://massgis.maps.arcgis.com/apps/OnePane/basicviewer/index.html?appid=47689963e7bb4007961676ad9fc56ae9#
> and then pick openstreetmap as the basemap. De-selecting "MassGIS
> Level 3 Assessors' Parcels" &  "MassGIS  Basemap Detailed Features"
> from the layers list also helps de-clutter things. If you check this
> then OSM buildings will be in light gray, and MassGIS buildings in
> dark gray.

Interesting that you can use osm in that but not in OLIVER.

> For most of the towns on the south shore I have checked, our buildings
> look reasonably good. However, Braintree and N Weymouth (north of
> route 3), have terribly misaligned or wrongly shaped buildings. Is
> there someone who maintains that sort of thing, or knows how to
> replace it with new geometry? It's not really a matter of new
> construction it just looks like the original data imported was bad.

That's an automated edit which needs a page/proposal.

I think that a few towns (near Cambridge) had buildings imported very
very early on.

Then, there was a "statewide building import" run by Jason Remillard,
and coordinated on this list (or its predecessor).  Lots of people did
quality checking.

So I wonder if you can point to a bad building or two and the changeset
when it was added.   If we have bad import data I think the first step
is to characterize where it is and how it got there.

> Someone who knows Boston may also want to check the new data for
> there. The MassGIS buildings page says, "In August 2019 MassGIS
> replaced the original Rolta-produced Boston structure polygons with
> those obtained from the City of Boston in March 2019. The city's data
> are more current and slightly more precise. MassGIS also added a few
> large structures built recently that were not in the Boston data."

That sounds tricky.  I think we will need to do some sort of automated
comparison.  One thing to be careful of is that when we imported
buildings it turned out that there were a few that either never existed
or did not, and people removed them.  So we have to make sure not to add
them back.


One thing I'm starting to try to understand is the relationship between
the MassGIS coordinates which I think are in NAD83(2011), and how we
transform them to OSM coordinates in "WGS84", which is a fuzzy term and
we just discussed on talk@ that that means "WGS84(curent)" which is
WGS84(G1672).  This is all very complicated and hard to follow if you
aren't into this as a hobby, but there can be about a 2m difference,
which I think now matters.


Greg