I like questions where the answer isn’t obvious yet.
I’m a financial economist turned data scientist. I build models, research systems and small products around markets, behaviour and uncertainty.
Can a public record tell us anything useful about tomorrow’s German power market?
Every forecast is sealed before the answer is known. Missed days stay visible. An AI assistant can explain the record, but it cannot rewrite it. The point is not to claim an edge early; it is to make the eventual answer harder to fool myself about.
The submitted snapshot’s running skill is descriptive only. Formal assessment remains pending until 31 October 2026.
Follow the experiment →
Different subjects. The same habit: make the uncertainty visible.
The homepage is the map. Each case study contains the methods, evidence, decision and boundary.
Forecasts, prices and decisions made before the answer arrives.
Tools that expose their sources, limits and human hand-offs.
Some work begins with a dataset. Some begins because I want the thing to exist.
A financial economist who became more interested in how the answer is made.
I was trained in markets, econometrics and uncertainty. Data science gave me a way to turn that training into working systems: find the real decision, choose a credible baseline, test what changes, and keep the boundary visible when the evidence runs out.