Hello OSM US and import community,
I am proposing a phased, human-reviewed project using the Oklahoma 9-1-1 Management
Authority’s public NG911 road-centerline data to reconcile local-road names and add
individually confirmed missing minor/local roads throughout Garvin County, Oklahoma.
The disclosed consultation scope is Garvin County, but editing would begin with a
ten-road pilot and pause for review. This is not a statewide upload, highway
replacement, automated geometry update, or deletion/replacement of existing OSM roads.
No OSM edits have been made. I am requesting feedback on the permission evidence,
scope, conflation method, tagging approach, and proposed review workflow before any
editing begins.
Documentation
- Proposer’s normal OSM account: GeoSafe-Moshe | OpenStreetMap
- Proposer’s OSM Wiki account: User:GeoSafe-Moshe - OpenStreetMap Wiki
- Dedicated import-only account: GeoSafe-Moshe_Import | OpenStreetMap
- Import wiki page: Oklahoma/Import/Garvin - OpenStreetMap Wiki
- Source catalog: https://okmaps.org/OGI/search.aspx
- Oklahoma NG911 standards:
https://okmaps.onenet.net/documents/OK_ADDRESS_STANDARDS_3_APPROVED_05-02-2024.pdf - Public code and prepared review sample:
oklahoma-ng911-osm-import/counties/garvin at main · GeoSafe/oklahoma-ng911-osm-import · GitHub
Permission
The Oklahoma 9-1-1 Management Authority provided this written response concerning
the public OKMaps data:
The centerline and map point data that we have on the public facing side of
OKMaps is open to anyone that wants to use it. The ESZ data is not for public
release.
This project uses only the public road-centerline layer. It does not use ESZ data.
I would like the import community’s view on whether this written statement is
sufficient for incorporation into OSM under the ODbL. The correspondence can be
provided privately to the appropriate OSM reviewer with personal email addresses
redacted from public materials.
Proposed attribution is Oklahoma 9-1-1 Management Authority on the OSM Contributors
wiki/import page and in the source description of relevant changesets.
Dataset and initial screening
The source is locally maintained NG911 road-centerline data aggregated by the state.
The evaluated Garvin records are associated with Garvin County 911,
psap.8274.ok.gov.
- 5,000 county-tagged source segments
- 9 expired records excluded
- 4,991 unique active segments screened
- 1,735 likely existing local-road matches, not proposed for import
- 576 numbered-highway segments excluded
- 324 segments excluded by a major OSM-road match
- 115 private, driveway, or placeholder-name segments excluded
- 56 same-name alignment/partial-gap cases separated for review
- 100 apparent missing local-road source segments, grouped into 66 named review units
The name audit is counted by both source rows and unique OSM ways; these diagnostics
are not additive. Among local-road matches, 1,776 source rows already have the same
primary name ignoring case and need no name change. The 2,100 rows needing primary-name
attention correspond to 945 unique OSM ways and 377 distinct NG911 FullName values.
Of the matched ways, 110 are associated with multiple NG911 names and are ambiguous.
A preliminary one-to-one screen identified 308 unnamed OSM ways associated with
exactly one NG911 name. For 240 of these ways, all matched source segments are within
5 m; 278 stay within 10 m. These are screening candidates, not approved edits, and
will be reduced further by relation, access, road-class, current-version, and visual
correspondence checks. Separately, 535 named OSM ways have one OSM primary name versus
one NG911 primary name; those conflicts require slower manual review.
The OSM comparison uses a checksum-verified Geofabrik Oklahoma extract containing
OSM data through 2026-08-30T20:21:06Z. It will be refreshed again immediately
before any approved editing.
Scope
The consultation covers Garvin County local-road name reconciliation and individually
confirmed missing public minor/local roads. Work will be executed in documented small
batches, beginning with the ten-road pilot. It excludes:
- Interstate, US, Oklahoma state, turnpike, ramp, and other major roads;
- existing OSM road geometry replacements;
- private roads, driveways, service drives, alleys, parking lots, and trails;
- address points, address ranges, and ESZ data;
- automatic conversion of NG911
RoadClassto OSMhighway=*; and - automatic or unattended uploading.
Conflation and review
The preparation tool compares normalized primary and alternate source names with OSM
names, alternate names, official names, local names, old names, and refs. It performs
nearby-geometry screening and recognizes common Oklahoma rural-grid variants such as
North County Road 3200 and N 3200 Road.
The NG911 FullName is treated as the authoritative local name for an approved
feature. Name normalization is only for detecting possible matches; it does not
select the output tag. Thus NORTH COUNTY ROAD 3258 would be written as
name=North County Road 3258, not shortened to a legacy TIGER form. A differing
name on a geometrically matching OSM way triggers manual conflation review rather
than creation of a duplicate. The existing way and history are preserved, and an
old value is retained as alt_name=* or old_name=* only when its continued or
historical use is supported.
The actual Garvin source contains 862 distinct FullName values. All are uppercase,
none contains periods, and none uses standard road-type abbreviations such as RD,
CR, ST, AVE, BLVD, or HWY. Output therefore changes letter case only; it
does not invent expansions. Abbreviation rules are used only to match legacy OSM or
TIGER names against the complete NG911 name.
For a proposed name-only addition to an existing unnamed OSM way, automated screening
will confirm one source name, close correspondence, local-road scope, and the absence
of detected conflicts. A reviewer will then confirm that the OSM way represents the
same complete road corridor. Only name=* is proposed; geometry and all existing tags,
nodes, relations, and history are preserved.
Missing-road and existing-name-conflict candidates receive the slower individual
QGIS/JOSM review. For missing geometry, reviewers must verify that the road:
- physically exists;
- is public and is not a driveway, track, or paper road;
- is not already represented under another name or with offset geometry;
- can be assigned an observable OSM functional road classification; and
- connects correctly at intersections and county boundaries.
Uncertain candidates will be rejected or deferred. Existing OSM objects, history,
relations, restrictions, and mapper improvements will be preserved.
Proposed tagging
There is no bulk field-to-tag conversion. For each approved missing road:
highway=residentialonly for an observable residential street;highway=unclassifiedonly for an observable minor public road serving the local
road network;name=*from the Oklahoma NG911FullName, with only normal OSM capitalization
and punctuation applied;- other tags such as
surface=*,access=*,oneway=*,bridge=*, orlayer=*
only when independently verifiable.
Internal NG911 IDs, PSAP/ESN fields, address ranges, and maintenance metadata will not
be copied to OSM object tags.
Proposed editing and changesets
The proposed phases are:
- upload, if approved, only the reviewed ten-road pilot and pause for community review;
- process high-confidence one-name/unnamed-way cases in geographically coherent
batches after additional automated and visual checks; - review existing-name conflicts individually, without automatic overwriting; and
- review the 66 apparent missing-road groups individually for existence, access,
classification, imagery, endpoints, and topology.
Ambiguous ways associated with multiple NG911 names will be deferred unless resolved
individually. Initial changesets will contain approximately 10–25 road-level changes;
the exact size will follow community feedback and remain small enough for meaningful
review and reversion.
The reviewed first pilot contains exactly ten road-level cases: three geometry cases
and seven name-only cases. All ten passed an initial user visual review on 2026-08-31.
The three geometry cases remain held for an independent topology, access,
classification, endpoint, and imagery check. The seven name-only cases are ready for
proposal review, but not for upload:
North County Road 3258— possible missing road;East County Road 1500— possible missing continuation adjoining OSM way 14763489;Alvera Lane— locally confirmed new extension adjoining OSM way 676767064;Butler Road— proposed name for existing unnamed OSM way 14762516;North County Road 3025— proposed name for existing unnamed OSM way 14762878;North County Road 3230— proposed name for existing unnamed OSM way 14761025;North County Road 3382— proposed name for existing unnamed OSM way 14762474;North County Road 3160— proposed name for existing unnamed OSM way 14767266;East County Road 1565— proposed name for existing unnamed OSM way 14765535; andEast County Road 1632— proposed name for existing unnamed OSM way 14760596.
Each feature would be examined individually again against current OSM before editing.
Changesets would be small and geographically coherent and would include:
source=Oklahoma 9-1-1 Management Authority; survey/approved imagery as applicablesource:date=2026-07-06import=yes- a link to the approved import wiki page
The dedicated import-only account GeoSafe-Moshe_Import will perform every edit under
this project. The project will pause after its first changeset for community review
before continuing.
QA and rollback
Before and after each changeset, JOSM validation and routing/topology checks will cover
duplicate geometry, disconnected endpoints, unintended crossings, county-border seams,
and invalid tags. Source NGUIDs, review decisions, OSM IDs, and changeset IDs will be
recorded outside OSM.
If a changeset is found to be materially wrong, work will stop and that changeset will
be reverted promptly in coordination with the community. No further batch will begin
until the issue is understood and the documented process is corrected.
Thank you for reviewing this proposal. In particular, feedback on the permission
statement and the conservative ten-road missing-road/name-only scope would be
appreciated.