Energy Demand & Load Forecasting Course

Lesson 1 • Energy Consumption Concepts and Terminology

Defines load, demand, energy, and power and distinguishes peak from average consumption. Establishes shared vocabulary used throughout the course.

Lesson 2 • Load Profiles and Temporal Patterns

Analyzes daily, weekly, and seasonal load shapes across customer segments. Introduces load duration curves as a core analytical tool.

Lesson 3 • Demand Drivers and Influencing Factors

Identifies economic, demographic, behavioural, and weather variables that shape consumption. Links each driver to measurable indicators used in later models.

Lesson 4 • Structure of the Energy System

Maps the supply chain from generation to end use and identifies where demand signals originate. Provides system-level context for forecasting decisions.

Lesson 1 • Regression Analysis Fundamentals

Covers simple and multiple linear regression for relating demand to explanatory variables. Builds the core modelling skill applied in short- and long-term forecasting.

Lesson 2 • Forecast Accuracy Metrics

Defines MAE, RMSE, MAPE, and bias and explains when each metric is appropriate. Establishes the evaluation standard used to compare all models in the course.

Lesson 3 • Time-Series Decomposition

Separates trend, seasonality, and irregular components from historical load series. Enables analysts to isolate structural signals from noise.

Lesson 4 • Descriptive Statistics for Energy Data

Applies measures of central tendency, dispersion, and distribution shape to consumption datasets. Grounds statistical reasoning in energy-specific examples.

Lesson 1 • Model Validation and Backtesting

Applies rolling-window and out-of-sample testing to assess short-term model reliability. Connects validation results to operational deployment decisions.

Lesson 2 • ARIMA and Seasonal ARIMA Models

Builds autoregressive integrated moving-average models for stationary and seasonal load series. Introduces Box-Jenkins methodology as a rigorous short-term forecasting framework.

Lesson 3 • Exponential Smoothing Techniques

Applies simple, double, and Holt-Winters smoothing to capture level, trend, and seasonality. Provides fast, low-complexity baselines for operational forecasting.

Lesson 4 • Calendar and Special-Event Effects

Encodes day-of-week, holidays, and special events as model inputs to correct systematic biases. Ensures forecasts reflect real-world scheduling patterns.

Lesson 1 • Weather Normalisation Methods

Adjusts historical consumption for atypical weather to reveal underlying demand trends. Produces normalised baselines essential for medium-term planning.

Lesson 2 • End-Use Saturation Modeling

Tracks appliance ownership rates and usage intensity to project residential and commercial demand. Bridges bottom-up device-level data with aggregate forecasts.

Lesson 3 • Trend Extrapolation and Curve Fitting

Applies linear, exponential, and S-curve models to extend historical demand trends. Provides simple yet robust methods for medium-term projections.

Lesson 4 • Forecast Uncertainty and Confidence Intervals

Quantifies forecast uncertainty using prediction intervals and scenario ranges. Prepares analysts to communicate risk to planning and finance stakeholders.

Lesson 1 • Top-Down Macroeconomic Approaches

Links aggregate energy intensity to GDP and structural economic change for rapid long-term projections. Complements bottom-up models with a macro-level consistency check.

Lesson 2 • Long-Term Demand Drivers and Scenarios

Identifies structural drivers including demographics, electrification, and policy shifts over long horizons. Introduces scenario analysis as the primary tool for managing deep uncertainty.

Lesson 3 • Bottom-Up End-Use Forecasting

Aggregates device-level energy use across all sectors to build a transparent, auditable long-term forecast. Enables granular sensitivity analysis by technology or sector.

Lesson 4 • Stress Testing and Sensitivity Analysis

Systematically varies key assumptions to identify forecast vulnerabilities and planning risks. Produces robust planning ranges rather than single-point projections.

Lesson 1 • Tree-Based Ensemble Methods

Applies gradient boosting and random forests to capture nonlinear demand patterns. Demonstrates strong out-of-sample performance with interpretable feature contributions.

Lesson 2 • Model Deployment and Monitoring

Covers productionising ML forecasting models, scheduling retraining, and detecting performance drift. Bridges the gap between model development and operational use.

Lesson 3 • Probabilistic ML Forecasting

Extends point forecasts to full predictive distributions using quantile regression and conformal prediction. Supports risk-aware operational and planning decisions.

Lesson 4 • Recurrent Neural Networks and LSTM

Uses sequence-to-sequence architectures to model temporal dependencies in load series. Particularly effective for multi-step ahead forecasting with complex seasonality.

Lesson 1 • Demand Response Programme Fundamentals

Defines curtailment, time-of-use, and direct-load-control programmes and their triggering conditions. Establishes the DSM landscape that forecasters must account for.

Lesson 2 • Distributed Energy Resources and Load Shape

Models the net-load impact of rooftop solar, battery storage, and EVs on system demand profiles. Prepares forecasters for the increasing complexity of modern load shapes.

Lesson 3 • Energy Efficiency Programme Impacts

Quantifies gross and net energy savings from efficiency programmes using deemed and metered approaches. Integrates verified savings into medium- and long-term demand forecasts.

Lesson 4 • Measuring Demand Response Impacts

Applies regression discontinuity, matching, and difference-in-differences methods to isolate DR load reductions. Produces defensible impact estimates for planning models.

Lesson 1 • Integrating Forecasts into Planning Processes

Embeds demand forecasts into capacity planning, procurement, and rate-setting workflows. Demonstrates how forecast outputs drive tangible infrastructure and financial decisions.

Lesson 2 • Forecast Hierarchy and Reconciliation

Aligns forecasts across spatial and temporal aggregation levels using coherent reconciliation methods. Prevents inconsistencies between system, regional, and customer-level projections.

Lesson 3 • Forecast Governance and Documentation

Establishes review cycles, assumption registers, and audit trails for organisational forecast processes. Ensures reproducibility, accountability, and regulatory defensibility.

Lesson 4 • Communicating Forecasts to Stakeholders

Designs visualisations and narratives that convey forecast uncertainty and key assumptions clearly. Bridges the gap between technical analysts and non-technical decision-makers.