GUI-Based Data Model Creation in Sigma (Drag-and-Drop Approach)

Introduction

Modern analytics platforms demand flexibility, speed, and governance at the same time. Sigma Computing enables both technical and non-technical users to work on the same analytical layer without duplication of effort. One of the most critical foundations in Sigma is data model creation, which directly impacts performance, accuracy, scalability, and user trust.

In the Sigma implementation approach, we leverage both SQL-based modeling and Sigma’s drag-and-drop interface. This hybrid method allows us to maintain strong data governance and correctness while also empowering business users to explore data independently. This article explains how we design, build, and validate data models in Sigma.

What is data models in Sigma

In Sigma, a Semantic layer data model that defines business-friendly columns, table relationships, and reusable metrics. These models sit directly on top of cloud data warehouses like BigQuery, Snowflake, and Redshift.

Key Characteristics:

1.  No Data Duplication: Queries reflect live data directly from the warehouse.
2. Flexibility: Built using either SQL or the UI.
3. Consistency: Models can be reused across multiple workbooks to enforce business logic.
4. Lineage Visibility: Sigma provides built-in lineage visibility for drag-and-drop data models, allowing users to clearly trace how metrics and columns are derived from source tables through transformations.

Sigma  Supports Two Data Modeling Approaches 

1. Business-Friendly Drag & Drop Data Modeling

This approach allows users to visually model data without writing SQL, while still maintaining enterprise-grade governance.

1. Select Base Table: Define the primary grain using fact tables such as GL Balances or AR Aging
2. Visual Joins: Sigma automatically detects join keys; users validate join type and cardinality
3. Calculated Columns: Create model-level logic using Excel-like formulas
   (e.g., If([JournalStatus] = 'P', [AccountedCredit], 0))

2. Complex SQL-Based sigma Data Modeling

This approach is designed for advanced use cases where precise control over logic, grain, and performance optimization is required.

1. Explicit Grain Definition: Use base CTEs to clearly define the data grain
2. Avoid Fan-Out: Apply window functions such as ROW_NUMBER() to control joins and duplicates
3. Performance Optimized: Push complex transformations to the warehouse for efficient execution

Impact: Business Value Delivered by Dataplatr + Sigma

By combining Sigma’s modern capabilities with Dataplatr’s implementation expertise, customers achieve:

1. Faster time-to-insight with reduced dependency on data teams
2. Consistent KPIs across all dashboards and users
3. Improved performance through optimized warehouse queries
4. Higher adoption due to familiar spreadsheet-style interaction
5. Future-ready analytics that scale with data growth

Dataplatr helps customers design the right modeling strategy, avoid common pitfalls, and unlock the full value of their cloud data platforms.

Conclusion

Sigma represents a shift from traditional BI to collaborative, warehouse-native analytics. When implemented correctly, it enables both governance and self-service at scale. Through our structured, problem-driven approach, Dataplatr helps organizations build a robust sigma data model using SQL and drag-and-drop, delivering trusted insights, faster decisions, and measurable business impact.

Contact us at: [email protected]  
For consultations or custom inquiries: https://dataplatr.com/contact-us  
Linkedin