Our analytics team at Mack Weldon provided internal business stakeholders with access to the data to create their own analysis. However, inconsistencies and lack of documentation in the models posed challenges for users in understanding and utilizing the data effectively. This led to a decrease in efficiency and collaboration within the data environment.
To address these issues, we conducted a comprehensive audit of the analytics infrastructure. We identified models with inadequate documentation, ensured consistency between the backend data tool and BI tool, cleaned up deprecated data sources and redundant models, and examined dependencies of frequently used downstream models such as fact line items, orders, and customers.
We leveraged dbt packages like dbt utils (star, surrogate) to standardize tables and apply SQL best practices for improved query performance. During the audit process, we followed the agreed-upon dbt styling guide to maintain consistency and formatting standards.
Our team also documented any hidden or unused fields in LookML to provide clarity to business users. We cleaned up deprecated configurations in the Stitch environment by referring to API documentation to understand existing models. The improved consistency and documentation of models have led to enhanced collaboration and efficiency within the data environment.