|
About
About Electricity Market Monitor
Electricity Market Monitor is a real-time electricity market monitoring platform focused on providing insights into electricity market dynamics. The platform visualizes live generation, demand, pricing data, and carbon intensity across different regions.
This tool is designed to help energy professionals, analysts, and enthusiasts understand market dynamics through interactive visualizations and historical data analysis.
Forecasting Methodology
The platform includes machine learning-powered price forecasting for the APXMID (EPEX Spot Market Index Price) in the Great Britain electricity market. The forecast provides a 24-hour ahead prediction of the Volume Weighted Average Price (VWAP) of short term power contracts in the GB electricity market.
Model Architecture
The forecasting model uses XGBoost (eXtreme Gradient Boosting), a commonly used gradient boosting framework for time-series data. Specifically, we employ the gblinear booster, which uses linear models as base learners, providing a good balance between interpretability and predictive performance for time-series price data.
Features
The model uses the following features to predict APXMID prices:
- Market Fundamentals: Total system generation and total system demand
- Temporal Features: Hour of day, minute of hour, day of week, and month of year to capture seasonal and daily patterns
- Autoregressive Features (Lag Variables): Historical price values at specific time intervals to capture temporal dependencies and price momentum
Autoregressive Features (Lag Variables)
Electricity prices exhibit strong temporal dependencies, meaning past prices are highly predictive of future prices. The model incorporates autoregressive features (lag variables) that capture these patterns:
- Daily Lag (48 periods): Price from 24 hours ago (same time yesterday) to capture daily patterns
- Weekly Lag (336 periods): Price from 7 days ago (same time last week) to capture weekly patterns
- Monthly Lag (1440 periods): Price from 30 days ago (same time last month) to capture longer-term seasonal patterns
Note: The model requires at least 30 days of historical price data to compute all lag features. Forecasts generated with insufficient historical data may have reduced accuracy, particularly for longer-term lags.
Note: Price forecasts are generated automatically every 30 minutes alongside market data updates. These forecasts should not be used for any critical decision-making, they are indicative only and subject to error.
Data Sources
Electricity Market Monitor integrates data from multiple sources to provide live and historical market information:
Great Britain Market
- Elexon BMRS: Real-time and historical generation, demand, and market pricing data for the GB electricity system
- NESO (National Energy System Operator): Carbon intensity data and system operational information
Disclaimer: While we strive to provide accurate and up-to-date information, we do not guarantee the correctness or accuracy of the data displayed on this platform. Data may be subject to delays, errors, or inconsistencies. This tool is intended for informational purposes only and should not be relied upon for critical decision-making.
This website represents personal views and analysis and does not reflect the views of any employer or affiliated organizations.
About the Developer
Electricity Market Monitor was created by Jack Greenwood, a Product Manager and Market Analyst specialising in trading and optimisation systems within the energy sector. With experience across Japan, Great Britain, and European power markets, this platform brings together real-time data and analytics to provide insights into dynamic electricity markets.
This project is developed and maintained independently as a personal initiative.