SAS Viya: Enterprise AI, Analytics, and Data Science Platform

Overview

SAS Viya is an enterprise artificial intelligence (AI), analytics, and data science platform that provides an integrated environment for the entire analytical lifecycle. It is designed to support a wide range of analytical workloads, from data preparation and exploration to advanced predictive modeling, machine learning, and real-time decisioning. The platform's cloud-native architecture allows for scalability and deployment across various cloud environments, including AWS, Azure, and Google Cloud, as well as on-premises infrastructure simultaneously. This flexibility enables organizations to manage their analytical processes within their preferred operational frameworks.

The platform is suitable for enterprises that require robust governance and lifecycle management for their AI and machine learning models. It offers capabilities for managing models from development through deployment and monitoring, addressing compliance requirements such as GDPR and HIPAA. SAS Viya integrates various analytical components, including SAS Visual Analytics for business intelligence, SAS Visual Data Mining and Machine Learning for model development, and SAS Model Manager for MLOps. This integration aims to streamline the collaboration between data scientists, business analysts, and IT operations teams.

Developers and data scientists can interact with SAS Viya using a variety of interfaces, including Python, R, Java, Lua, and a comprehensive REST API. This multi-language support facilitates integration into existing data science workflows and allows users to leverage familiar programming environments. The platform also includes CASL (SAS Cloud Analytic Services Language), a proprietary language for interacting with the in-memory processing engine. SAS Viya is particularly beneficial for organizations that need to operationalize AI models at scale, manage complex analytical pipelines, and ensure the reliability and interpretability of their deployed models. Its emphasis on governance and MLOps capabilities distinguishes it in environments where regulatory compliance and model accountability are critical. For example, the platform's ability to integrate with existing enterprise data warehouses and data lakes, as described in Databricks documentation on SAS Viya integration, highlights its role in broader data ecosystems.

Key features

• Advanced Analytics and Machine Learning: Provides a suite of algorithms for predictive modeling, forecasting, optimization, and deep learning, supporting various analytical techniques.
• Data Management and Preparation: Offers tools for data ingestion, transformation, cleansing, and feature engineering, enabling users to prepare data for analysis.
• Cloud-Native Architecture: Built to run on Kubernetes, allowing for scalable deployment on public or private clouds and on-premises environments.
• MLOps and Model Governance: Includes capabilities for model registration, versioning, performance monitoring, retraining, and explainability, ensuring models are managed throughout their lifecycle.
• Open API and SDK Support: Facilitates integration with other systems and allows data scientists to work in Python, R, Java, and other languages via SDKs and REST APIs.
• Visual Analytics and Reporting: Provides interactive dashboards and reporting tools for data exploration and visualization, enabling business users to derive insights.
• Real-time Decisioning and Event Stream Processing: Supports processing data streams in real-time to detect anomalies, trigger alerts, and make immediate decisions.

Pricing

SAS Viya implements a custom enterprise pricing model. Specific costs are determined based on an organization's individual requirements, including the scope of deployment, number of users, and specific analytical components utilized. Organizations interested in deploying SAS Viya typically engage directly with SAS sales to obtain a tailored quote.

For detailed pricing inquiries, refer to the SAS Viya pricing page.

Common integrations

• Cloud Platforms: Integrates with AWS, Microsoft Azure, and Google Cloud for deployment and data services.
• Data Lakes and Warehouses: Connects to various data sources including Hadoop, Apache Spark, Snowflake, and Databricks for data ingestion and processing.
• Programming Languages: Offers SDKs for Python (SAS Scripting Wrapper for Analytics Transfer), R, Java, and Lua for custom application development and integration.
• Business Intelligence Tools: Can integrate with third-party BI tools through its reporting and API capabilities.

Alternatives

• Databricks: A unified data analytics platform offering data warehousing, data engineering, and machine learning capabilities on a lakehouse architecture.
• Amazon SageMaker: A cloud-based machine learning service that provides tools for building, training, and deploying ML models.
• TIBCO Data Science: An analytics platform focused on visual data science, model building, and operationalization across various data sources.

Getting started

To interact with SAS Viya programmatically, you can use the SAS Scripting Wrapper for Analytics Transfer (SWAT) package for Python. This example demonstrates connecting to a CAS (Cloud Analytic Services) server and loading a simple dataset.

import swat

# Replace with your CAS server host, port, username, and password
# For SAS Viya for Learners, these details are typically provided in your environment.
cas_host = 'your_cas_server_host.sas.com'
cas_port = 5570
cas_username = 'your_username'
cas_password = 'your_password'

try:
    # Connect to the CAS server
    conn = swat.CAS(hostname=cas_host, port=cas_port, username=cas_username, password=cas_password)
    print(f"Successfully connected to CAS server: {cas_host}:{cas_port}")

    # Create a simple in-memory table
    data = [
        ['A', 10, 100.5],
        ['B', 20, 200.0],
        ['C', 30, 150.75]
    ]
    columns = ['Category', 'Value1', 'Value2']
    cas_table = conn.addTable(name='my_test_table', data=data, caslib='casuser', replace=True, columns=columns)

    print("\nTable 'my_test_table' created in casuser caslib:")
    # List tables in the casuser caslib
    tables_info = conn.table.tableInfo(caslib='casuser')
    print(tables_info)

    # Fetch data from the table
    fetched_data = cas_table.fetch()
    print("\nFetched data from 'my_test_table':")
    print(fetched_data)

    # Perform a simple action, e.g., get descriptive statistics
    summary_stats = cas_table.summary()
    print("\nSummary statistics for 'my_test_table':")
    print(summary_stats)

except swat.SWATError as e:
    print(f"Error connecting or interacting with CAS server: {e}")
except Exception as e:
    print(f"An unexpected error occurred: {e}")
finally:
    # Disconnect from the CAS server
    if 'conn' in locals() and conn.connected:
        conn.disconnect()
        print("\nDisconnected from CAS server.")

This Python script establishes a connection to a SAS Cloud Analytic Services (CAS) server, creates an in-memory table, fetches its contents, and performs a basic summary statistics operation. This demonstrates a fundamental interaction for data scientists using Python to leverage SAS Viya's analytical capabilities.