Measure regional building and address gaps before adding datasets #408

Before you start

• Searched current issues and discussions; no equivalent task was found. Related work is linked below.

Objective

Establish whether extra building/address datasets add useful coverage or precision in selected regions before committing to ingestion or a new renderer.

Scope

Compare one well-mapped German urban/rural sample and one independently justified gap region with pinned existing basemap/geocoder inputs. Evaluate bounded candidate extracts, such as Overture buildings/addresses and relevant OpenAddresses sources, only after confirming actual coverage and per-source rights. Distinguish 2D footprints, heights/roof geometry and searchable address points. This task is measurement and a go/no-go, not planet ingestion, automatic OSM editing or merging the archived LoD2 prototype.

Verified current state

Rechecked against main at a8e03bc3efc49207d91f7c1855402c6cf149b8fd on October 7, 2026.

• 3D buildings layer already extrudes basemap building features; absence of a separate buildings ingest does not mean buildings are absent from the map.
• Pelias data build currently builds OSM and Who's On First inputs; optional sources are already scoped in Expand Pelias with optional address, GeoNames, and interpolation sources #300.
• Overture pull and ingest support the existing places pipeline, not a committed buildings/address-theme rollout.
• Primary candidate documentation: Overture buildings, Overture addresses, OpenAddresses. Overture addresses are currently documented as Alpha, with partial German coverage and source-specific licenses; recheck the selected release during the pilot.

Completion criteria

• Record baseline providers, exact data versions, bounding boxes and source rights; do not infer missing source data from a hidden style layer.
• Measure marginal valid footprints/address points, positional precision, duplicates/conflicts, address recall and useful height/roof detail with sampled independent checks.
• Identify shared upstream sources so recycled OSM/OpenAddresses data is not counted as entirely new coverage.
• Report region-specific freshness, build/update cost, tile/query overhead and storage, with a threshold for useful benefit agreed before evaluating.
• Check licenses/attribution and production rights for the exact extracts/services; an aggregate project license is insufficient for independently licensed address sources.
• Publish before/after evidence and separate decisions for footprints, address search and detailed roofs. Continue source implementation in existing Expand Pelias with optional address, GeoNames, and interpolation sources #300/Revisit German LoD2 buildings: archived prototype and production blockers #366 where scope matches; create only distinct follow-ups if a measured gap warrants them.
• Update documentation with coverage/rights limitations if a candidate is adopted. A no-go or adequate-existing-coverage result is valid completion.

Verification approach

Use versioned bounded extracts and independently sampled ground truth; compare counts only after spatial/identity deduplication. Include map-style-controlled screenshots, geocoder fixtures with caches isolated and a reproducible cost/coverage report.

Dependencies

#393 supplies repeatable German comparison locations. #300 already tracks optional Pelias address/GeoNames/interpolation sources; #366 tracks the archived German LoD2 prototype and production blockers. Neither should be duplicated by a new source-implementation issue. This measurement task can run independently of the main discovery/navigation sequence.

This is backlog planning, not authorization to expand the implementation beyond this scope. Resolve named decisions during triage before scheduling work.