About Giraffe Trendz

Who runs this

Giraffe Trendz is built and run by one person: an AI/ML engineer who has been collecting Pokémon cards since childhood and never quite stopped. By day I build machine-learning systems; on my time off I point the same tools at the Pokémon TCG market — because the hobby deserves better data than gut feel and hype threads.

That day job spans more than a decade of data and AI engineering at some of the biggest names in entertainment. I have built for the world's top streaming platforms — Disney+, HBO Max, Hulu, Paramount+, Peacock, and more — where the dashboards I built are the ones executives actually make decisions with. I've shipped AI-driven insight systems on LLM orchestration frameworks that turn raw data into trend analysis automatically, built statistical-inference pipelines that measure whether a product change really worked, and led and trained analytics teams along the way. Before streaming, I cut my teeth alongside the pioneers of search-era analytics — machine learning, predictive modeling, and web-scale data systems, back when those disciplines were being invented.

• 
• 
• 
• 
• 

Every one of those habits runs this site: measured numbers over vibes, experiments over opinions, and pipelines that do the work every night whether anyone is watching or not.

How the numbers are gathered

Every night the site records the market price of every Pokémon single and sealed product it tracks — more than two hundred thousand items, each condition separately — and files it as that day's snapshot. The weekly report compares the newest snapshot with the one from seven days earlier and ranks what moved by dollars, so a card that gained two dollars on a three-dollar base never outranks a card that gained two hundred.

Set and artist figures use the median of their members, which keeps one chase card from dragging a whole set's number around. Streaks require an unbroken run of daily rises, not a single spike. When a move traces to a single sale or a relisting rather than broad demand, the report says so next to the number.

What the site does not do: it does not scrape auctions, it does not track individual sellers, and it does not use any account data in the reporting. The methodology is the same one I would use at work, applied to a market I happen to love.

What the site does today

Giraffe Trendz tracks the Pokémon TCG market — hundreds of thousands of singles and sealed products — and turns it into the reporting you see here: the weekly movers article, set and artist trends, streak detection, and per-single price histories with every condition tracked separately.

Everything published is measured, not vibes. The article is generated from the numbers themselves, and where the data is thin — a single sale swinging a price, a coverage gap — the site says so instead of hiding it.

What's coming: predictive insights

The reason this site exists is where it's headed: predictive models for the TCG market. Which singles are building momentum, which spikes look like demand versus one-off sales, where a set's trajectory points next — the kinds of questions machine learning is actually good at, applied to a market I care about.

The models are in development and not live yet. I would rather ship them late than ship them wrong: they'll appear here once they hold up against real out-of-sample data, clearly labeled as predictions with their track record shown. Nothing on the site today is a prediction — it is all measurement of what already happened.

The ground rules

This is market reporting and, eventually, model-driven insight — never financial advice. Prices come from public marketplace data; Giraffe Trendz is independent and not affiliated with any marketplace, The Pokémon Company, or Nintendo. The full details live in the Terms of Use and Privacy Policy.

Say hello

Questions, corrections, or a card you think the data is wrong about: grassisgreengg@gmail.com. If you want the data in your own hands, create a free account and open the Analyze workbench.