From energy markets to models that show their working.

I am Varun Rout, an applied data scientist based in Birmingham, UK. Most of my professional work has been forecasting: residual-load and price-curve modelling at E.ON, network-charge risk modelling before that, and multi-horizon demand forecasting at Manor Park Trading Company, the only role that carried the title “Data Scientist”.

The E.ON contract was fixed-term and ran its course at the end of 2025. Since January 2026 I have put that time into an independent portfolio, with a deliberate focus on football analytics: an opponent-adjusted expected-goals model benchmarked directly against StatsBomb’s own metric, a causal uplift model for retail campaign targeting, and a betting-market backtest that is honest about not beating the market.

The thread across all of it is the same: a model is not finished when it trains. It is finished when you know what it gets wrong, what it should be compared to, and what you would not claim about it. That is what this site tries to show on every project page, not just the ones with good numbers.