Founder — Fishing Intelligence

FTMO Trader Challenge Certificate awarded to Wytse Punter on 24 February 2021 — FTMO Trader Challenge passed, 24 February 2021

MEET THE FOUNDER

Hi, I'm Wytse Punter.

Predictive modeling has been my thing since 2020. It started with algorithmic trading, and in early 2021 I got funded by FTMO. That was the moment statistical edges stopped feeling theoretical. By day I work as a Data Engineer at Bentacera, and the rest of my time goes to markets, water, and the models that connect them.

Between the two worlds I kept noticing something odd. Trading has hundreds of datasets, dashboards, and models you can build conviction on. Fishing? Mostly gut-feeling tips, no real data backbone. That gap stayed with me.

VisScanner is where those worlds meet: the same techniques I used to find patterns in market data, now applied to catch records, water quality and biodiversity across the Netherlands.

WHY I BUILT THIS

It started with bread from a balcony.

Walking past a canal one day, I saw someone throwing bread from their balcony into the water. It triggered an idea: if this happens often, especially at apartment blocks near the water, those spots should be exceptional for fish. A constant stream of food pulls them in.

So I went looking for open-source data on water quality. To my surprise a huge amount was available, including fish populations. I started importing the catch records, and accidentally pulled in data on everything else too: aquatic plants, insects, plankton, anything you can think of.

I've always loved data engineering, building predictive models, and fishing, but I never imagined there'd be this much data until that bread story. I dove in with six years of algorithmic trading and data engineering behind me. Unlike in financial markets, the statistical edges here were lying around for the taking.

On 6 April 2026, an early version of the model flagged a day with 1.8× carp activity near Westzaan. Joshua Rijnhart went, trusting the forecast. First bite, landed. That moment turned a hobby project into something that actually works for other people.

THE APPROACH

Open data. Honest validation.

The stack is deliberately transparent: open data from Rijkswaterstaat, KNMI, GBIF and Kadaster (PDOK), scikit-learn for the prediction models, and every forecast is validated against catches the model has never seen before.

No black box, no marketing claims. If a model can't hold up on unseen data it doesn't ship, same rule as the trading models.

6 years in algorithmic trading

6+ open datasets combined

100% validated on unseen data

ON THE WATER

Half the work starts here.

Every model prediction eventually meets reality on a bank somewhere. That's how I find out which features actually matter, and every catch becomes another datapoint.

These photos span roughly fifteen years. I've been fishing since I was a kid. The data intuition came later, but the water always came first.