Project Overview
One page, plain language: what this platform is, who uses it, and how an event
runs from setup to hall of fame.
**Visual version with the full timeline and role tabs:
0 ** · Full spec: SRS · Step-by-step: WORKFLOWS/
What this is
The FLA Event & Awards Platform — a complete system for running public-voted, AI-assessed award events:
- 70% public votes + 30% board/judge score (weights configurable per event)
- Admin veto — absolute, audited, applies before or after results
- AI-conducted applicant interviews with per-category random questions and instant marking
- LinkedIn-verified voters, one reasoned vote each, verifiable receipts, hash-chained votes
- Winner lock as a single transaction — final standings, certificate numbers and
finalists can never drift out of sync
- Ceremony-day tools — stage screen, MC dashboard, award-team shields, self check-in
- Everything audited — every meaningful action recorded with actor, IP and metadata
Who uses it
| Role | What they do |
|---|---|
| Admin | Branding, events, approvals, vetoes, winner lock, system health |
| Staff | Day-to-day: registrations, manual entry, comms, tickets, check-in |
| Applicant | Apply → AI interview → track status & votes → share card |
| Voter | LinkedIn sign-in → reasoned vote → receipt → verify |
| Judge | Rubric review of assigned candidates (30% weight) |
| MC | Full-screen ceremony control: slides, winner queue, call-ons |
| Award team | Confirms winners are present before announcement ("shield") |
| Auditor | Read-only oversight of numbers and the audit log |
The journey in 10 steps
- Event setup — categories, question bank, branding, weights, toggles (all DB-driven)
- Registration opens — visitor (QR ticket) / award applicant (form + photo) / guest of honor; nominations too
- AI interview — randomized questions per category, answers marked, status → AI-scored
- Approval & screening — staff approve; duplicate detection and waitlists enforced
- Public voting — LinkedIn-verified, reason required, receipts, hash chain
- Judge review — rubric scores normalized into the board weight
- Veto & integrity — vetoes applied and audited; vote chain verified
- Winner lock — one transaction: standings + cert numbers + finalists published
- Ceremony day — stage screen, MC view, shields, self check-in
- After — share cards, certificates, hall of fame, season report, audit export
How scoring works
final = 70% × normalized(public votes) + 30% × normalized(board rubric)- One person, one vote (per candidate or per event) — duplicates blocked, rate-limited
- A veto removes a candidate from winners and public results entirely
- Results are a stored snapshot — public pages and DB always agree
Under the hood
Plain PHP 8.3 (no framework) + MariaDB (29 idempotent migrations, transactions on every critical write) + AI gateway (MiniMax provider with a keyless mock mode) + GD share cards + file cache + PWA/web push. Verified by a 34-check surface suite and a 266-check end-to-end suite that exercises the entire lifecycle.
New in v3.4.0
- Post-event feedback survey —
/survey(1–5 ratings + comment for attendees of the latest finished event). - Printable name badges —
/badges(admin/MC/award-team; each badge shows name, category, certificate number, and personal/self-checkin?t=TOKENlink). - Data-retention jobs —
/admin/retention(roll up page views, archive old audit rows, clean old notifications). - One-click backup — "Run backup now" button on
/admin/system. - Transparency hub —
/about-scoring(detailed scoring, veto, hash-chain, lock explanation).