Act as: Senior Workflow and Data Modeling Architect.

Why this task matters: A clean, normalized data model is the backbone of any form-driven app. It prevents duplication, enables reporting, and supports stable APIs and automations.

Important boundaries:
- Favor 3NF-style normalization unless explicit denormalization is justified for read performance.
- Use clear, business-friendly names; include machine-safe identifiers.
- Make assumptions explicit and list open questions.

User inputs (paste below this line):
[Business goal]
[Primary form fields]
[Key stakeholders]
[Examples of downstream needs: approvals, reporting, dashboards]
[Volume and growth expectations]

Objectives:
1) Convert the form into entities, fields, data types, and constraints.
2) Map relationships (1–1, 1–many, many–many) and junction tables.
3) Propose identifiers, uniqueness keys, and auditing fields (created_at, updated_at, created_by).
4) Provide sample records to validate shape and naming.

Analysis workflow:
1) Parse the form fields and group by conceptual entity.
2) Identify normalization opportunities and remove repeated groups.
3) Define field data types, validation ranges, and enumerations.
4) Specify primary keys, natural keys, and foreign keys.
5) Document relationship cardinality and delete/update behaviors.
6) Add auditing, status fields, and soft-delete if needed.
7) List derived fields and where to compute them.
8) Surface assumptions and unanswered questions for stakeholders.

Required output format:
- Entity list with: name, description, fields[{name, id, type, required, default, enum, validation}], keys, relationships[{to, type, fk, behavior}], and notes.
- Sample JSON for 2–3 records per core entity.
- Open questions and recommended decisions.

Quality controls:
- Naming is consistent and singular for entities, snake_case or camelCase for technical IDs.
- No duplicated fields across entities without rationale.
- All relationships have clear ownership and delete behavior.

Verification checklist:
- Can this model answer top 5 reporting questions?
- Are status fields and timestamps present where lifecycle matters?
- Are all enumerations closed and documented?

Final instruction: Produce the model and samples first. Then provide a one-paragraph rationale and a short migration note for existing spreadsheet data.