What is Semantic Modeling Language (SML)?

Understanding SML

The Semantic Modeling Language (SML) is an open-source, YAML-based language designed to define and manage semantic models. As a universal standard, SML enables different platforms to share semantic models, fostering portability and collaboration. By describing data in business-friendly terms, it allows users to query datasets using familiar language, bridging the gap between traditional Business Intelligence (BI) and next-generation Artificial Intelligence (AI) applications.

SML translates raw data into meaningful, actionable business terms that users across the organization can easily understand. Metrics in SML summarize data, such as “Sales” or “Count of Customers”. The SML captures what tables and columns contain the needed data and how to summarize it. Sales could combine data from “web_sales”, “catalog_sales” and “store_sales” tables. SML also specifies dimensions (e.g., fiscal year or region), which are attributes used to aggregate and slice the data in meaningful ways. Users combine metrics and dimensions through Business Intelligence (BI) tools or AI models using SML to get a single, consistent answer to their questions.

For example, consider a scenario where one user asks an AI interface for “Sales for the last fiscal year for Canada.” In contrast, another user creates a report in a BI tool using the same attributes. Since both interfaces use the same underlying semantic layer defined by SML, both users will receive the same answer regardless of how complex the underlying logic may be.

SML’s core goal is to democratize data access by providing a framework that can be understood and used by both technical and non-technical users. This accessibility is particularly significant as organizations increasingly rely on advanced analytics and AI for decision-making. SML sets the foundation for scalable, efficient, and transparent analytics workflows by enabling consistent and reusable models.

How SML Works

SML’s design is object-oriented, prioritizing composability and inheritance. This structure allows semantic objects—such as metrics, dimensions, and calculations—to be shared and reused within and across organizations. This capability reduces redundancy and ensures consistency across data models. Below are the key aspects of how SML functions:

• Familiar Syntax: SML employs YAML, a widely adopted, human-readable syntax that is easier for users with diverse technical backgrounds to learn and implement. YAML’s straightforward structure ensures that semantic models are both readable and writable.
• CI/CD Compatibility: SML integrates seamlessly with Git for version control, enabling Continuous Integration and Continuous Deployment (CI/CD) workflows. This ensures that semantic models can be efficiently developed, deployed, and maintained throughout their lifecycle.
• Comprehensive Coverage: SML supports advanced constructs, including multidimensional models, hierarchies, semi-additive measures, many-to-many relationships, and cell-based expressions. These features make it suitable for complex data landscapes and diverse use cases.
• Extensibility: SML’s syntax can be customized to accommodate specific organizational needs. This flexibility allows organizations to incorporate unique business logic and terminology into their semantic models.
• Direct Integration: SML connects directly to data platforms, eliminating the need for data extraction or movement. This enables real-time analytics and reduces latency in data processing.

Benefits of SML

Adopting SML offers many benefits, addressing critical data and analytics management challenges. Here’s how SML drives value:

1. Standardization: SML establishes a universal framework for semantic modeling, reducing the fragmentation caused by proprietary tools. This standardization simplifies collaboration and interoperability across platforms.
2. Interoperability: As a “Rosetta Stone” for analytics, SML enables the seamless translation of semantic models between tools like Tableau, Power BI, Looker, and Excel. This interoperability enhances flexibility and reduces barriers to adopting new technologies.
3. Reusability: SML’s object-oriented design facilitates the creation of reusable components, such as shared metrics and dimensions. This reduces duplication and simplifies the maintenance of semantic models.
4. Freedom from Vendor Lock-In: SML’s open-source nature allows organizations to migrate between platforms without rebuilding semantic models, fostering greater agility and reducing operational costs.
5. Enhanced Collaboration: SML offers a common language that bridges the gap between technical teams and business users. This collaboration ensures semantic models align with business needs while maintaining technical precision.
6. Accelerated Insights: SML’s ability to standardize business logic enables faster time-to-insight, empowering organizations to make informed decisions more quickly.
7. AI and BI Integration: SML provides the structured framework necessary for accurate natural language queries and other AI-driven applications. It ensures consistency in responses and improves the usability of AI for business intelligence.
8. Data Democratization: SML empowers non-technical users to engage with data more effectively by providing predefined metrics and intuitive query capabilities. This democratization enhances decision-making across all levels of an organization.

SML and the Role of Semantic Layers

Semantic layers play a critical role in modern data analytics by bridging the gap between raw data and actionable insights. SML amplifies the effectiveness of semantic layers through its standardized, open, and extensible approach. Here are the ways SML enhances semantic layers:

• Standardization and Portability: By establishing a universal modeling standard, SML ensures semantic models are easily shareable and portable across platforms. This eliminates the need to rebuild models when migrating to new tools, saving time and resources.
• Interoperability: SML decouples semantics from specific tools, enabling organizations to integrate multiple analytics platforms seamlessly. This interoperability eliminates silos and promotes real-time analytics without the need for data movement.
• Collaboration and Reusability: SML’s object-oriented design supports cross-functional collaboration by allowing both technical and business users to contribute to model development. Its reusable components ensure consistency across projects and teams.
• AI and BI Integration: SML enhances both traditional BI and AI applications by providing a structured framework. For instance, it ensures that AI models generate accurate, context-aware responses by grounding them in consistent semantic definitions.
• Governance and Compliance: SML enables robust data governance by defining clear access controls and lineage tracking. This ensures compliance with regulations while maintaining data security and integrity.

Best Practices for Using SML

To maximize the potential of SML and ensure successful implementation, organizations should follow these best practices:

1. Leverage CI/CD Workflows: Store semantic models in Git repositories to enable version control and streamlined deployments. CI/CD workflows facilitate collaboration and ensure changes are thoroughly tested before deployment.
2. Design Modular Models: Use SML’s object-oriented capabilities to create modular, reusable components that can be shared across teams and projects.
3. Engage Cross-Functional Teams: Involve both technical experts and business stakeholders in the model development process. This ensures that semantic models meet organizational needs while maintaining technical accuracy.
4. Focus on Governance: Define clear access rules and ensure transparency through lineage tracking. This helps maintain trust in data and ensures compliance with regulations.
5. Adopt Pre-Built Models: To accelerate implementation and reduce development effort, take advantage of pre-built semantic models available in cloud marketplaces.
6. Tailor Models to Business Needs: Customize SML syntax to reflect your organization’s unique terminology and processes. This ensures semantic models align with specific business goals.
7. Prioritize Training: To ensure the widespread adoption and effective use of SML, provide training for both technical and non-technical users.

Empowering Open Source Semantics

The open-source foundation of SML is transforming the analytics landscape by promoting collaboration, innovation, and interoperability. Here’s how SML is driving a new era in data and analytics:

• Standardization: By providing a universal modeling language, SML simplifies the integration of diverse analytics tools and reduces fragmentation in the analytics ecosystem.
• Community Innovation: SML’s open-source nature invites contributions from a global community of developers and data professionals. This collaborative approach ensures that SML evolves to meet its users’ needs.
• Reduced Vendor Lock-In: Organizations can adopt SML without being tied to a specific vendor, enabling greater flexibility and adaptability to future technologies.
• Accelerated Analytics: Pre-built semantic models and the ability to translate between platforms enable organizations to generate insights faster, empowering more agile decision-making.
• AI-Driven Applications: SML provides the semantic foundation for AI to generate accurate, context-aware responses. This makes it critical for deploying generative AI in real-world business scenarios.
• Democratized Data Access: By making semantic models accessible to non-technical users, SML promotes data-driven decision-making at all levels of an organization.

AtScale has played a pivotal role in the development and proliferation of SML, driving its vision to become a universal standard for semantic modeling. By open-sourcing SML, AtScale aims to foster a collaborative ecosystem where organizations and individuals can build, share, and innovate on semantic models without barriers. AtScale’s leadership in this space has been instrumental in bridging the gap between traditional BI and AI capabilities, ensuring that businesses can leverage the full potential of their data.

By integrating SML into its platform, AtScale empowers organizations to achieve consistent and governed insights across analytics tools while promoting interoperability and scalability. This commitment to open-source semantics underscores AtScale’s dedication to advancing data democratization and enhancing the analytics landscape.

What’s Next for SML

The Semantic Modeling Language (SML) significantly advances how organizations define, manage, and share semantic models. Its open-source, YAML-based design ensures accessibility, standardization, and interoperability across platforms and industries. By bridging the traditional BI and AI gap, SML empowers organizations to harness their data’s full potential.

A robust and flexible semantic layer becomes increasingly critical as businesses adopt advanced analytics and AI. SML’s ability to democratize data access, reduce vendor lock-in, and support both technical and business users positions it as a cornerstone of modern analytics. By embracing SML, organizations can achieve greater agility, scalability, and collaboration in an ever-evolving data landscape. To learn more about SML, check out our GitHub repository, which contains pre-built semantic models incorporating standard data models such as TPC-DS, common training models such as Worldwide Importers and AdventureWorks, and marketplace models such as Snowplow and CRISP.