LookML

LookML is the modeling language used by Looker, a business intelligence and data analytics platform. It enables data teams to define the semantics, calculations, and relationships of data sources in a centralized, reusable way. LookML models are written as a collection of model files and view files that describe how data from a database is exposed to end users through Explores, dashboards, and reports.

Core concepts in LookML include views, models, and Explores. A view file describes a database table or derived table and lists the fields it exposes. Fields are organized into dimensions and measures; dimensions capture attributes, while measures define aggregations such as sum, average, or count. A model file defines Explores and how they join to other views, establishing the relationships between data sources. Derived tables and persistent derived tables (PDTs) allow users to materialize subqueries for performance and reuse.

LookML relies on SQL generation. When a user runs a query, Looker composes SQL from the LookML definitions and sends it to the connected database. This enables centralized logic for filtering, aggregations, and calculations across dashboards and reports. LookML also supports dynamic behavior through user attributes and access filters to tailor queries per user and enforce data-level security.

Development and collaboration are supported through Git-based version control. LookML projects are typically stored in Git repositories, with Looker’s development mode and web-based IDE facilitating collaboration, review, and deployment across environments.

History and ownership: LookML was created for Looker, which was acquired by Google in 2019 and integrated into Google Cloud. Today, LookML remains the core modeling language for Looker and its data modeling workflow.