pokemon-value-predictor

A deployment-ready MVP for predicting undervalued Pokémon TCG cards across 30-day, 90-day, and 6-month horizons using transparent rule-based scoring.

Monorepo structure

• /frontend — React + Vite + Tailwind dashboard.
• /worker — Cloudflare Worker API + D1 scoring backend.
• /worker/migrations — D1 schema migrations.
• /worker/data — mock seed data for local bootstrap.
• /wrangler.toml.example — Cloudflare deployment template.

Scoring model (rule-based)

upside_score =

• momentum_score * 0.25
• catalyst_score * 0.30
• grading_score * 0.20
• volume_score * 0.15
• sentiment_score * 0.10

Confidence buckets:

• high >= 75
• medium >= 55 and < 75
• low < 55

Backend endpoints

• GET /api/cards
• GET /api/predictions
• GET /api/predictions/top?horizon=90d&limit=20
• GET /api/cards/:id
• POST /api/ingest/mock-prices
• POST /api/score/run

Includes a scheduled handler as a cron placeholder for daily scoring workflows.

Quickstart

1) Frontend

cd frontend
npm install
npm run dev

2) Worker + D1 local

cd worker
npm install
cp ../wrangler.toml.example ../wrangler.toml
npm run db:migrate
npm run db:seed
npm run dev

MVP notes

• Uses only mock/sample data (no paid APIs required).
• Code structure is prepared for future provider integrations (TCGPlayer, eBay, PSA, Reddit, Discord) via ingest jobs and adapters.
• Optional R2 bucket binding included for archived snapshots.

Suggested next steps

1. Add provider-specific ingest adapters under worker/src/providers/*.
2. Add simple auth for admin-only POST endpoints.
3. Persist scored snapshots to R2 for historical strategy comparisons.
4. Replace mock sentiment inputs with derived Reddit/Discord metrics.