Contributing campus data

Campus-data changes belong in GapwiseHQ/data, not in a downstream consumer.

Entrance contributions

For a missing or incorrect UTM entrance, use the visual Gapwise Data contributor at data.gapwise.ca/contribute.

You can choose a building, click the doorway on the map, drag the pin to refine its position, and record only the access, direction, or barrier-free facts you actually know. On a phone, the editor can optionally use browser location while you are surveying an entrance. Building-specific links such as https://data.gapwise.ca/contribute?building=MN open with that building already selected.

A public submission is review evidence, not an automatic mutation of canonical routing data. Unknown is a valid value, and submitted claims stay subject to validation and maintainer review before they can become canonical facts.

Good contributions

Useful contributions include:

• missing or corrected entrances submitted through the visual contributor;
• corrected building names, aliases, or stable codes;
• source-backed coordinates or footprint improvements;
• verified entrance/accessibility evidence;
• routing graph corrections;
• provenance metadata and source identifiers;
• tests or validators that catch invalid campus facts;
• documentation of uncertainty or known limitations.

Maintainer and code workflow

Use a Data repository pull request when the change affects canonical files, validators, schemas, data-production tooling, or other maintainer-owned implementation rather than a field observation that fits the visual contributor.

1. Change the canonical data or data-production tooling in GapwiseHQ/data.
2. Run the repository validator and integrity checks.
3. Let the core-consumer contract verify that current Gapwise can still compile and route against the candidate dataset.
4. Merge Data.
5. Sync/pin the validated snapshot into core through the supported synchronization path.
6. Let normal core CI validate product behavior before production deployment.

Do not make equivalent campus-fact edits independently in web, mobile, AI, or documentation repositories.

Evidence expectations

Prefer official/source-linked evidence when available. If a useful value is inferred, mark it as inferred rather than upgrading it to verified simply because it appears plausible on the map. For field observations, use the visual contribution form’s Not sure options instead of guessing.

For questions about how an artifact is represented or distributed, see Dataset catalog and Distribution and versioning.