A daily market intelligence pipeline for the Australian jigsaw puzzle market, built by Premium Puzzles.
The Index treats the puzzle market the way a quant treats an equity market: titles are constituents, brands are sectors, and demand signals are prices. It answers questions no supplier or competitor can:
- Is overall puzzle demand growing or shrinking right now?
- Which brands are gaining share, and which are bleeding it?
- Is this market broad-based, or is it an S&P 500 run by a Magnificent Seven of hit titles?
- Which specific puzzles are breaking out before the trade catches on?
Scrapers collect public demand signals daily across retail, search, and community platforms. The pipeline aggregates them into a composite score per title, brand sub-indices, and a headline Puzzle Market Index level. The scoring methodology and weights are the proprietary part.
scrapers/ -> data/raw/ (local only, never committed)
pipeline/ -> data/processed/index.json (the published output)
web/ -> interactive visualisations embedded on the website
The only data published from this repo is the derived index in
data/processed/index.json. Raw source data stays local.
Collection layer (four sources) and the Index scoring engine are built and
running. python run_daily.py collects today's due sources and rebuilds
data/processed/index.json. A daily local task + a GitHub Actions cron keep
it fresh (see docs/OPERATIONS.md). The web visualisation is the remaining
piece. Model: docs/INDEX_DESIGN.md. Hard-won scraping knowledge:
docs/SCRAPING_LEARNINGS.md.