UNFIRE POWER / RESEARCH
AI4CLIMATE · 2026 DEMO

Make flexible energy use easier to schedule.

An experimental decision aid for Spanish households and small businesses. It uses a simple trained forecasting model to rank electricity-use windows, explains the trade-off, and leaves every decision with the person.

Research-only demonstration. All hourly prices and historical examples below are generated synthetically for reproducibility. They are NOT live Spanish tariffs, grid readings, or validated emissions data. This demo makes no claim of actual CO₂ reduction, customer adoption, or appliance integration.

Illustrative 24-hour price forecast

€/kWh · synthetic
Earliest-window cost
—
Suggested-window cost
—
Illustrative difference
—
Predicted hourly price Suggested window

Model: 5-nearest-neighbours trained on 9 simulated days of hourly prices, evaluated on one held-out simulated day. This is an experimental machine-learning price forecaster, not a validated operational model. Its purpose is to show a testable recommendation workflow.

Plan a flexible task

SUGGESTION · REQUIRES HUMAN APPROVAL
Calculating…

No equipment or account is controlled.

✓ Approved locally as a hypothetical plan. No device command or booking has been executed.