1. Purpose
Hawaiʻi's Green Fee asks visitors to help fund the stewardship of the places they come to experience. That funding relationship carries a transparency obligation in both directions: the public and the Legislature need to see where the money goes, and the visitor deserves to re-encounter what their fee built. This toolkit demonstrates a registry-driven publication architecture that meets that obligation without creating a new dashboard platform for the State to buy, host, or maintain — and without turning stewardship awardees into dashboard administrators.
2. Design objectives
- Preserve State ownership of project information and control of its GIS dashboards.
- Reduce reporting burden on awardees — they contribute stories, photos, highlights, and field observations, nothing more.
- Support future funding cohorts without dashboard redevelopment.
- Establish project-level provenance for every public claim.
- Publish through a human-held gate — a person publishes, never an automated process.
- Carry place-based stewardship context (moku) alongside, never in place of, official program data.
3. The architecture in six layers
| Layer | Platform | Owns |
|---|---|---|
| State authority | State program administration | Projects, awards, contracts, official outcomes — the authoritative record, always |
| Registry governance | Version-controlled repository | Award Registry Manifest: identity, cohort membership, publication eligibility, routes |
| Automation | CI pipeline | Registry update → validation → manifest build → site build → publication; the pipeline is the audit trail |
| Public portal | Static site (this demonstration) | Program overview, project directory, outcome summaries, stewardship storytelling |
| Geospatial experience | State ArcGIS | Dashboards, StoryMaps, feature services — the portal links out, it never replaces |
| Executive analytics | Power BI | Internal review and cross-agency reporting from published exports |
The pattern at the seam: portal page → ArcGIS dashboard. Registry records carry the State's dashboard identifiers when the State registers them; geospatial authority never moves.
4. Replication framework
A new funding cohort is one file: FY2027.json, validated against the same schema,
published through the same gate, rendered by the same build. Replicating the architecture for a
different program — or a different jurisdiction — is the same controlled method the demonstration
itself followed:
- Define the business question the public surface must answer.
- Validate the local data against its authoritative source.
- Map terminology and workflow to the jurisdiction's own vocabulary.
- Approve the metric definitions with the program owner before anything renders.
- Test quality — schema validation plus registry invariants in CI.
- Review with subject-matter experts before the publication gate opens.
- Publish with limitations stated — on every page, in every export.
5. The FY2026 demonstration cohort as case study
The sample cohort holds six fictional records across five moku and four islands, deliberately split four published / two held, so the architecture's central controls are visible in the running system rather than described in prose: the publication gate (held records are counted but build no pages and export in no feed), the release bar (a record cannot publish without a story, outcomes, and provenance), the provenance chain on every project page, and the machine-readable published-manifest projection that releases exactly what the pages show.
The cohort also demonstrates the closing of the fee's loop: its reef-stewardship sample pairs resource monitoring with shoreline visitor briefings — the visitor who paid the fee meeting the steward the fee funds, at the place itself, with that encounter captured as a reported metric.
6. Stewardship context: the moku
Every record reads in its moku — the traditional land division whose community the project serves. Moku are a regional axis (whose place), carried alongside and never in place of the State's regulatory classifications (what may happen where). The registry claims no land-use authority; the moku context is what lets a project page speak to its community and its visitors in terms of place.
7. Boundaries
All records in this demonstration are fictional samples; the v0 schema structurally requires the sample marking, so real program data cannot validate against it. State systems remain authoritative for every fact this architecture publishes. Nothing here decides funding, eligibility, policy, or enforcement.