[Imports] Racine County Building Footprints w/ Address
Joshua Carlson
jdcarls2 at gmail.com
Sat Oct 20 16:09:41 UTC 2018
Hello!
My name is Josh Carlson. I've been on OSM longer than my account suggests,
but I've only really been taking it seriously for about the past year or so.
The current focus of my OSM editing has ballooned out into a full-blown
import in scope, so I am pausing my efforts to go through the proper
guidelines. I haven't yet made a wiki page or dedicated OSM username, but
here are the proposed details of the import:
1. Goals
1. Map all building footprints in Racine County
2. Include accurate address tags
2. Schedule
1. No real timeframe, though I'd hope to have it finished within 6
months or so.
3. Import Data
1. Microsoft Computer-Generated Building Footprints
1. url: https://github.com/Microsoft/USBuildingFootprints
2. license: ODbL
3. notes: A good starting point, but neither accurate nor thorough
enough to be used as-is. As noted below, will be
reviewed/refined before
import.
2. Wisconsin V4 Parcel Layer
1. url: https://www.sco.wisc.edu/parcels/data/
2. license: No license is specified. However, the WI Parcel
Initiative, as laid out in Act 20, is explicitly intended for
general use
in both the public and private sectors, and is free to
acquire, use, and
distribute. Is something more "official" required here?
3. notes: In the V4 Observation Reports (
https://doa.wi.gov/DIR/V4_Observation_Reports_Statewide.pdf),
Racine County is known to have a small number of minor issues in the
address fields I'm interested in. These are very small relative to the
total number of addresses, but worth noting.
3. Import Type
1. one-time
2. not automated, as building footprints often need minor
corrections, and smaller buildings are often missed.
4. Data Preparation
1. Reduction / Simplification
1. Clip WI building footprints geojson to Racine County boundaries
2. Spatially join footprints to underlying parcel layer
1. Opt for "wholly within" predicate to avoid address mistakes;
will have to review buildings which straddle parcel lines
as they are
identified.
3. Remove all fields not related to site address
2. Tagging Plan
1. Create new fields which appropriately concatenate site address
fields; new fields will conform to addr:housenumber,
addr:street, addr:city
and addr:postcode conventions
3. Data Transformation
1. Using QGIS to perform joins and field calculations
2. Split layer out into separate files for each PSLL Qsection,
possibly QQsection in dense urban areas.
4. Data Transformation results
1. Prior to now, I've been working straight from the shapefiles,
merging buildings one by one as I review them in JOSM. once things are
re-worked as XML files, I can post a link on a wiki page
5. Data Merge Workflow
1. Doing this solo, unless anyone wants to jump in on it.
2. Workflow
1. Open a building footprint layer
2. Using DG Premium Imagery (at highest level of detail), adjust
and add building footprints
3. Merge into "final" *.osm file to await upload.
3. Changeset size policy
1. Changesets will likely be uploaded in PLSS sections or quarters
sections, depending on the density of features
6. Conflation
1. Since I've been reviewing all footprints individually, I've just
been handling conflation issues as they arise. Racine's buildings are not
well-mapped, so this does not happen often. When buildings are already in
the OSM dataset, they often still lack addr:* tags, which I will
copy over.
So, thoughts? This is obviously still in the early stages of planning.
Community feedback has been invaluable to me, and I gladly welcome
suggestions on this.
Thanks!
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openstreetmap.org/pipermail/imports/attachments/20181020/f4be3edd/attachment-0001.html>
More information about the Imports
mailing list