AI/ML ModelOps is a Journey. Get Ready with SAS® Viya® Platform on Azure

This post has been republished via RSS; it originally appeared at: New blog articles in Microsoft Community Hub.

Why do you need a ModelOps Platform for your organization? 

 

If you are a data scientist or an analytics leader, you know the challenges of developing and deploying analytical models in a fast-paced and competitive business environment. You may have hundreds or thousands of models in various stages of the analytics life cycle, but only a fraction of them are actually delivering value to your organization. You may face issues such as long development cycles, manual processes, lack of visibility, poor performance, and loss of intellectual property. These issues can prevent you from realizing the full potential of your analytics investments and hinder your ability to innovate and respond to changing needs.

 

That’s why you need model ops in your organization. Model ops is a set of practices and technologies that enable you to automate, monitor, and manage your analytical models throughout their life cycle. Model ops can help you streamline your workflows, improve your model quality, increase your productivity, and ensure your models are always aligned with your business goals. Model ops can also help you foster collaboration and trust among your stakeholders, such as data scientists, IT, business users, and regulators. With model ops, you can turn your models into assets that drive value and competitive advantage for your organization.

 

SAS Viya Platform: A powerful AI/ML Model management Platform

 

SAS Viya is a powerful cloud-based analytics platform built by Microsoft's coveted partner SAS Institute Inc. that combines AI (Artificial Intelligence) and traditional analytics capabilities. SAS Viya seamlessly integrates with Microsoft Azure services, enhancing the analytics capabilities and providing a powerful platform for data-driven decision-making.

 

Simplified Acquisition Process for SAS Viya Platform @Azure Marketplace

 

The simplified acquisition process of SAS Viya on Azure marks a significant departure from the traditional method of obtaining SAS Viya. With the introduction of SAS Viya on Azure, users now benefit from a streamlined approach via the Azure Marketplace. This marketplace serves as a user-friendly platform where, with just a click on the "Create" button, customers initiate the acquisition process.

The automated deployment, a key feature of this approach eliminating the need for extensive IT involvement. Azure Marketplace not only expedites the transaction offering a faster and more accessible path for users to access and leverage the advanced analytics capabilities of SAS Viya.

 

Key Features of SAS Viya on Azure:

 

  1. SAS Visual Machine Learning (VML): SAS Viya includes robust machine learning capabilities. With VML, you can build, train, and deploy machine learning models efficiently. Whether you’re a Python enthusiast, an R aficionado, or prefer Jupyter Notebooks, SAS Viya integrates seamlessly with these languages to enhance your data science workflows. 

  2. Integration with Python, R, and Jupyter Notebooks: SAS Viya provides native integration with popular programming languages. You can leverage your existing Python or R code within the SAS Viya environment. Additionally, Jupyter Notebooks allow for interactive exploration and documentation of your analyses.

  3. Azure Data Sources and Services Integration:

    • Azure Synapse: SAS Viya connects to Azure Synapse, enabling you to enrich data from various sources. Use SAS Information Catalog and SAS Visual Analytics to explore and prepare data efficiently.
    • Azure Machine Learning: Collaborate between SAS Viya and Azure Machine Learning to build and deploy analytic models. You can choose your preferred programming language and even use a visual drag-and-drop interface for model components. Seamlessly move models to production.
    • Power Automate and Power Apps: Automate decision-making processes by integrating SAS Viya with Microsoft Power Automate and Power Apps. Enable real-time, calculated decisions across various domains, such as claims processing, credit decisioning, fraud detection, and more.
    • Azure IoT Hub and Azure IoT Edge: Stream data from IoT devices across environments, allowing real-time decisioning and analysis. Benefit from in-stream automation, data curation, and modeling while maintaining total governance.
    • Azure Data Sources: SAS Viya uses high-performance connectors to source data from Azure environments, provisioning data for downstream AI needs.

 

SAS Analytic Lifecycle Capabilities

 

  1. Access and Prepare Data:

    • SAS Viya allows you to handle complex and large datasets efficiently. You can perform data preparation tasks such as cleaning, transforming, and structuring data for analysis.
    • Whether your data resides in databases, spreadsheets, or cloud storage, SAS Viya provides seamless connectivity to various data sources.
  2. Visualize Data:

    • Data visualization is crucial for understanding data relationships, patterns, and trends. SAS Viya offers powerful visualization tools to create insightful charts, graphs, and dashboards.
    • Explore your data visually, identify outliers, and gain valuable insights before diving into modeling.
  3. Build Models:

    • SAS Viya leverages AI techniques to build predictive and prescriptive models. Whether you’re solving real-world business problems or conducting research, you can utilize machine learning algorithms, statistical methods, and optimization techniques.
    • Experiment with different models, evaluate their performance, and choose the best one for your specific use case.
  4. Automation:

    • Automate repetitive tasks within your analytic workflows. SAS Viya allows you to create data pipelines, schedule data refreshes, and automate model deployment.
    • Collaborate with other users by sharing workflows and automating decision-making processes.
  5. Integration:

    • Connect seamlessly with open-source languages such as Python and R. If you have existing code or libraries, integrate them into your SAS Viya environment.
    • Leverage the power of both SAS and open-source tools to enhance your analytics capabilities.
  6. ModelOps:

    • Managing models over time is critical for maintaining their accuracy and relevance. SAS Viya provides tools for monitoring model performance, adapting to changes in data, and retraining models as needed.
    • Stay on top of your models’ health and ensure they continue to deliver value.

 

Demonstration of Deployment on Azure:

 

Let’s walk through the steps for deploying SAS Viya on Microsoft Azure. This process typically takes about an hour to complete. Here’s how you can get started:

  1. Access SAS Viya on Azure:

  2. Deployment Form:

    • Click “Create” next to the plan.
    • Complete the form with the necessary information:
      • Project Details:
        • Specify your Subscription and Resource Group. These values depend on your organization’s Azure resource management practices. (Don't forget to prepare a suitable landing zone for your Analytical platform)
      • Instance Details:
        • Personalize the URL that users will use to access SAS Viya. The Region and Deployment DNS Prefix contribute to the URL. Choose a region geographically close to your users.
        • Provide a Deployment Name for reference within the Azure interface.
      • Security and Access:
        • Set an administrator password for SAS Viya (remember or record it).
        • Choose one of the following options for SSH public key source:
          • Generate a new key pair.
          • Use a key stored in Azure.
          • Copy and paste a public key.
        • Optionally, secure access by specifying Authorized IP Ranges.
  3. Review and Create:

    • Click “Review + create” to proceed.
    • Confirm the information, accept the terms and conditions, and click “Create.”
    • If you opted to create SSH keys, select “Download + create” when prompted.
  4. Deployment Completion:

Summary

 

  • SAS Viya on Azure provides a user-friendly, automated approach to access advanced analytics capabilities.
  • Offers seamless integration with Azure tools and services, catering to a wide range of users in the analytics space.

Next Steps 

 

In the next part, I will dive deep into the model ops capabilities and advantages of combining Azure and SAS Viya platform. 

Leave a Reply

Your email address will not be published. Required fields are marked *

*

This site uses Akismet to reduce spam. Learn how your comment data is processed.