SAS Energy Forecasting Cloud
From generation to distribution, get repeatable, traceable and defensible energy forecasts in the cloud. Scale up and down depending on the requirements of your business.
SAS Energy Forecasting Cloud dramatically reduces the time, resources and effort traditionally needed to produce accurate, defensible and scalable forecasts by automating 70% of the training, scoring, model selection and forecasting process. This 'as a service' offering allows utilities and smart cities to generate more efficient and accurate load forecasts, including those for renewable generation, without having to host and maintain software."María Luisa Paradinas Head of Business Development Innova
Key features
Automation, scalability, statistical sophistication and transparency for operating more efficiently and effectively at all levels of decision making.
Multihorizon energy forecasting
Supports short-, medium- and long-term forecasting for electric loads and renewable generation – meeting the needs of operations, market participation and strategic planning across all timeframes.
Automated model generation & ranking
Automatically generates, evaluates and ranks candidate models using a combination of statistical, machine learning and hybrid algorithms, selecting the best-performing model per use case.
Hierarchical forecasting & reconciliation
Performs forecasting at multiple levels (e.g., meter, feeder, substation, region) and ensures alignment between granular and aggregated forecasts through automated reconciliation.
Advanced driver variable selection
Dynamically identifies the most predictive weather, calendar and exogenous variables, using automated testing and statistical validation to maximize forecast accuracy.
Scenario forecasting
Allows users to create and compare planning scenarios – such as economic variables – for risk evaluation and strategic decision making.
Calendar & event sensitivity modeling
Captures the effects of holidays, weekends, time changes and special events on energy use or production through built-in, calendar-aware modeling capabilities.
Forecast accuracy monitoring
Allows tracking of forecast performance using key metrics (e.g., MAPE, RMSE, MAE) – enabling proactive model maintenance.
Web-based forecast management interface
Offers a user-friendly portal for running forecasts, viewing results, managing runs and downloading results – all without code or advanced training.
Batch CSV file integration
Enables input and output through standardized CSV files, supporting compatibility with existing IT systems, manual workflows and data preprocessing pipelines.
Secure & scalable cloud deployment
Built on a cloud-native architecture that ensures elastic scalability, performance monitoring and secure access control.