SAS Visual Forecasting
Plan better for the future with a configurable AI-enabled, end-to-end forecasting system you can trust.
Key features
SAS Visual Forecasting provides an open forecasting ecosystem for quickly and automatically producing a large number of reliable forecasts.
Data access, preparation & quality
Access, profile, cleanse, transform and manage your data using an intuitive interface that provides self-service data preparation and governance capabilities with embedded best practices and automations.
Data visualization
Visually explore data and create and share smart visualizations and interactive reports through a single, self-service interface. Perform goal seeking and scenario analysis in a fast and simple way. Augmented analytics and advanced capabilities accelerate insights and help you uncover stories hidden in your data.
Automatic time series analysis & forecasting
Analyze your time series in depth using automated capabilities or function packages, and easily convert time-stamped transactional data into a time series format to generate forecast models automatically.
Machine learning & deep learning
Incorporate machine learning (ML), neural networks (NNs), recurrent neural networks (RNNs) and hybrid techniques (NNs + time series) in the forecasting process to model even the most challenging series. Automatically generate features and transform transactional data into the right format to eliminate manual, labor-intensive feature engineering prior to using these sophisticated techniques.
Hierarchical modeling & reconciliation
Develop customizable pipelines to enable in-depth analysis and modeling at each hierarchical level. Then your forecasts will be automatically reconciled across all levels, by apportioning the changes as needed to maintain consistency throughout the hierarchy.
Open source integration
Integrate and parallelize Python or R code and algorithms to run in the cloud using a well-governed, consistent framework. Empower open source users to scale their code to run fast in a distributed manner using SAS Viya’s worker nodes in the cloud. Easily reuse open source forecasting algorithms in all business areas by creating custom nodes you can embed in your forecasting pipelines and share with colleagues.
Time-series segmentation
Automatically segment data based on demand-classification attributes, such as volume and volatility, using a prebuilt template to model each segment separately in the project pipelines. Or import custom-defined segments based on your business knowledge. Apply the most appropriate forecasting technique based on the nature of the data to improve forecast quality significantly.
Events management
Model the effect of events such as holidays, retail promotions, and natural disasters on dependent time series to improve model accuracy. The solution includes default prebuilt events like major holidays and an intuitive UI for developing custom ones.
Interactive & ensemble modeling
Analyze individual time series, compare models visually and develop custom models for individual time series via a simple user interface. Assess and evaluate models with automatically generated diagnostic plots and tables, and select your own model champions. Develop forecasting pipelines using different modeling strategies and use ensemble modeling to always select the best-performing model for each series.
Highly flexible forecast override
Make customized adjustments to forecast values using a powerful manual override capability via a simple UI. Define specific filters to select groups of time series defined by attributes that are not limited to just the hierarchical variable.
Additional forecasting & econometrics procedures
Address virtually any forecasting and time series analysis challenge with access to SAS/ETS® and the SAS Forecast Server procedures.
Distributed, accessible, cloud-native
Runs on SAS Viya’s fast, scalable and distributed in-memory engine. Call SAS actions and procedures from SAS, Python, R and Java. Use public REST APIs to add SAS data and AI capabilities to other applications or vice versa.