Multiple users can explore data, then interactively create and refine descriptive and predictive models. Distributed, in-memory processing slashes model development time, quickly surfacing valuable insights you can act on.
Uncover opportunities faster than your competitors.
Your data scientists and statisticians can act on observations at a granular level using the most appropriate analytical modeling techniques. The result? You'll unearth insights at unprecedented speeds, and find new ways to grow revenue.
Put better models into action faster.
Easily build and refine models to target specific groups or segments, and run numerous scenarios simultaneously. You can ask more what-if questions to get better results. And put results into action with automatically generated score code.
Boost analytical productivity.
Empower multiple users to interact with data visually – to add or change variables, remove outliers, etc. Instantly see how changes affect your model's predictive power, and make refinements quickly. You can also access SAS analytical algorithms from other environments using the programming language you prefer – Python, Java, R or Lua.
- Visual data exploration and discovery. Use SAS Visual Analytics to easily identify predictive drivers among multiple exploratory variables, and visually identify outliers and data discrepancies.
- Descriptive modeling. Visually explore and evaluate segments for further analysis using k-means clustering, scatter plots and detailed summary statistics.
- Predictive modeling. Build predictive models using techniques like linear regression, logistic regression, generalized linear models and decision trees – with point-and-click ease.
- Open, code-based model development. Programmatically access analytical actions from SAS Studio, call them from other languages – Python, R, Lua, Java – or use public REST APIs to add SAS Analytics to existing applications.
- Dynamic group-by processing. Concurrently build models and process results for each group or segment without having to sort or index data each time.
- Model comparison and assessment. Generate model comparison summaries (lift charts, ROC charts, concordance statistics, misclassification tables) on one or more models.
- Model scoring. Export models as SAS DATA step code, and apply them to new data.
- Distributed, in-memory analytical processing. Build models faster on diverse data sets, including Hadoop. There's no need to shuffle data or write data to disk.
- Flexible deployment options. Runs on commodity hardware, or in a private or public cloud infrastructure.