Insights

Delivering Data for Analytics

This is the third in a short series of articles where Jake Wedge, Senior Product Engineer, explores how modern data platforms can be built in a faster, more consistent and more maintainable way using DataOps.

The Gold layer is where data becomes consumable for reporting, dashboards and analytics. Whilst Data Vault provides a scalable foundation for integrating and governing data in the Silver layer, business users need data organised into structures that are easy to query and understand. This is where dimensional modelling comes in.

In this article, we'll explore how Seriös ONE supports Kimball-style dimensional modelling through a metadata-driven approach, combining automation for common patterns with the flexibility of Spark SQL.

Gold Layer: Kimball with SQL Flexibility

The Gold layer is where business users live. They need data organised into familiar star schemas: fact tables with measures and dimension tables with descriptive attributes. Seriös ONE Core makes this transformation straightforward whilst giving data engineers complete freedom to express complex business logic.

Here's the crucial design choice: whilst the framework provides standardised handling for Kimball patterns (surrogate keys, SCD types, default rows, incremental merges), you define transformations using Spark SQL. This gives you the full expressiveness of SQL for complex joins, window functions, conditional logic, and aggregations.

Continue reading to explore how Seriös ONE handles dimension and fact tables, automates common Kimball patterns and uses metadata to support lineage and AI across the Gold layer.