# Student project brief
## Testing a provenance-first model for public-land works records

**For:** capstone, honours or postgraduate projects in GIS and spatial science, environmental science, environmental health, planning, information management or environmental governance.
**Materials:** the Substriva Research & Education sandbox and its synthetic sample dataset (CSV and GeoJSON). All records are synthetic. No council, agency or university participation, endorsement or ethics approval is implied by this brief.

### Background
Public land is managed by many organisations: councils, state agencies, water authorities and committees of management. Records of the works done on that land are scattered across systems, staff and contractors. Substriva is testing a model in which each record keeps its source, custodian, date precision, review state and publication state. Anything the evidence doesn't establish stays **UNKNOWN**.

### Possible project tasks (choose one or combine)
1. **Evaluate the metadata model.** Is the schema enough for defensible provenance? Compare it with an established standard (e.g. W3C PROV, ISO 19115 lineage) and propose additions.
2. **Evaluate the treatment of UNKNOWN.** How should missing or partial evidence be shown to different audiences? Design and run a small comprehension test with peers (keep it to classmates unless your institution's ethics process approves more).
3. **Test the spatial representation.** Assess the 25 m public generalisation, point versus polygon records, and how boundary uncertainty should be shown.
4. **Assess usability.** Run a heuristic evaluation of the sandbox walkthrough and report the five most important issues.
5. **Propose interoperability improvements.** How could records from two different custodians be read together without merging authority? Sketch the data contract.
6. **Assess the research value of temporal works records.** For one environmental or public-health question of your choice, what would a works history need to contain to be useful? What could it never establish?

### Deliverables (suggested)
- A short report (2,000–6,000 words, depending on level) with methods, findings and limitations.
- Any proposed schema or design changes, as a diff or mock-up.
- A one-page summary for a non-specialist land manager.

### Ground rules
- Use synthetic data only, unless a land manager separately and formally agrees otherwise.
- No claims of safety, treatment status, clearance or causation.
- Any work with human participants follows your institution's ethics requirements. This brief does not provide ethics approval.
- Substriva may use anonymised findings to improve the model, with your agreement. Agree intellectual property and publication terms with your supervisor before starting.

### Contact
Use "Propose a student project" on the Research & Education page.