# GridLoom — European Energy Forecast Lab

> Precomputed European demand, solar and price forecast experiments with held-out backtests. Not live forecasts, certified market predictions or network power-flow simulation.

- Canonical URL: https://khanix.co.technology/
- Methodology & Equations: https://khanix.co.technology/methodology/
- Interactive Browser Runtime: https://khanix.co.technology/run/
- Author & Root Entity: Mohammad Rezwan Khan, PhD (https://mrkhan.co.technology/#person)
- Ecosystem Directory: https://mklab.co.technology/
- Knowledge Graph JSON-LD: https://khanix.co.technology/knowledge-graph.jsonld

## Governing Mathematical Formulation & Units
- Formula: `ŷ_{t+h} = β_0 + Σ_k β_k Φ_calendar(t+h) + Σ_j α_j y_{t+h−24j} (Calendar-Ridge + Local-Seasonal ensemble with empirical residual quantile intervals [q_{0.10}, q_{0.90}])`
- Units: Load and solar generation in megawatts (MW); day-ahead electricity price in EUR/MWh; history window = 56 days (1,344 hourly rows ending 2026-10-05T00:00:00Z); forecast horizon = 72 hours.

## Reproducible Worked Example
Market DE:DE-LU (Germany/Luxembourg bidding zone, 1,344 hourly training rows from 2026-08-10 to 2026-10-04 UTC): At 2026-10-05T00:00:00Z, the 72-hour ahead load forecast is 40,682.27 MW with empirical interval [38,991.57 MW, 42,372.98 MW], evaluated alongside solar generation (MW) and price (EUR/MWh) held-out backtests.