Day-ahead electricity demand forecast
Inspired by publicly reported work at NESO
Forecast tomorrow's electricity demand half-hour by half-hour from past demand and the calendar.
The brief
NESO, which runs Great Britain's electricity system (formerly National Grid ESO), has written about using machine learning to forecast demand and publishes its data openly. Using its historic demand data, forecast the next day half-hour by half-hour, compare with simple baselines, and see where forecasts fail (holidays, cold snaps). Good looks like: MAPE against a same-day-last-week baseline, error by season and holiday, and a clear forecast chart.
Read the case first
What NESO has said in public about this kind of work, on its own pages:
- ESO and The Alan Turing Institute use machine learning to help balance the GB electricity grid (NESO) ↗
- NESO Data Portal ↗
Make it your own
Forecast for India: a state's demand from Grid-India reports, or your college's electricity use from its meter readings, with Indian holidays and heat waves.
AITeky has no link with NESO, which has not reviewed or endorsed this brief. The brief only draws on what the company has published, uses open data rather than its data, and is a learning exercise: don't present your project as NESO's work or system.
What you will build
- Download historic demand data
- Explore daily, weekly and yearly patterns
- Build a same-day-last-week baseline
- Train a gradient-boosted model with calendar features
- Analyse errors on holidays and extreme days
Free, open data to build it with
Sign up free to get the data and links
It takes a minute with your email. AiTeky Pass (₹999/month or ₹9,999/year) is free for your first 30 days.