- All Foundries
- 商品管理と取引
- 顧客とマーケティング戦略
- デジタルエンジニアリング
- ステーション・スタジオ
- サプライチェーン＆インダストリー5.0
- サステナビリティ
- Physical AI
- Next-Gen Infrastructure
Gas & Electricity Demand Forecasting & Scenario Simulation
The initiative moved forecasting from fragmented model preparation toward a more repeatable business workflow. By bringing together gas and electricity demand signals, portfolio assumptions and scenario logic, the organization gained a stronger foundation for planning decisions in a volatile energy environment.
Forecasting had to reflect the real operating context of an energy retailer, where demand is shaped by external conditions, portfolio dynamics, aggregation levels and distinct gas and electricity modelling needs.
The initiative needed to account for weather, calendar effects, portfolio changes and different aggregation levels without oversimplifying the drivers of consumption.
Business planning required models that could manage multiple explanatory variables and time horizons while staying connected to operational forecasting needs.
Commercial and planning teams needed to reduce dependence on manual scenario preparation, making assumptions easier to reproduce, compare and refresh.
Electricity forecasting required hourly models by access tariff and telemetered supply point, including observed and unobserved supplies.
NTT DATA connected historical data, explanatory variables, forecasting models and scenario simulation into an operational workflow that could be refreshed as new information became available.
Quantified key demand drivers, including calendar and weather variables, so forecasts reflected the factors shaping gas and electricity consumption.
Built short- and medium-term gas-demand models at daily and monthly levels, across different aggregation levels.
Created an agile simulator using portfolio, unit-consumption and temperature assumptions, enabling teams to test and recover future scenarios.
Developed electricity-demand models by access tariff and telemetered supply point, incorporating observed and unobserved supplies.
The initiative helped lower exposure to energy-purchase deviation penalties, with approximately EUR 1 million in documented gas deviation penalty savings. The weighted daily gas forecast error was reduced by more than 4 points, and the electricity average daily forecast error was reduced to below 2%. Beyond model accuracy, the capability gave commercial and planning teams a faster, more repeatable basis for scenario preparation, budgeting and demand-related decision support.
