Looker

Google's enterprise business intelligence platform — used by data and analytics teams to build self-serve dashboards and reports on top of cloud data warehouses.

Looker is a cloud-based business intelligence and data exploration platform acquired by Google in 2020 and integrated into Google Cloud. It uses a proprietary modeling language called LookML to define metrics and data relationships in a semantic layer, enabling business users to explore data without writing SQL. Looker is deployed by mid-to-large companies running data warehouses on BigQuery, Snowflake, or Redshift, and appears frequently in data analyst, analytics engineer, and data engineer job postings. Looker Studio (formerly Data Studio) is the free, simpler counterpart for dashboard creation.

Typical time to job-readiness: ~3 weeks.

Learning Looker

Beginner

Learn Looker as an end user: navigate Explores, build looks and dashboards, understand dimensions vs measures, and use filters and pivots. Looker's own interactive training (Looker Learning) is the best free starting point for both users and developers.

Intermediate

LookML fundamentals: views, models, explores, dimensions, measures, and derived tables. Understand how LookML maps to SQL and how to build a semantic layer that non-technical stakeholders can self-serve without asking for SQL queries.

Advanced

Advanced LookML patterns — persistent derived tables (PDTs), symmetric aggregates for fanout correction, liquid templating for dynamic dimensions, and Looker API for embedding and automation. Analytics engineering interviews often include a LookML modeling exercise alongside dbt.

Key concepts

  • LookML — Looker's modeling language that defines dimensions, measures, and data relationships as code
  • Explore — the query interface where users select dimensions and measures; generated from LookML model definitions
  • Dimension vs measure: dimensions are attributes (customer name, date), measures are aggregations (count, sum)
  • Semantic layer — Looker's core value: define business logic once in LookML, expose it consistently to all users
  • Looks vs dashboards: a Look is a single saved query/visualization; a dashboard combines multiple Looks
  • Derived tables — SQL queries defined within LookML to create virtual tables for complex transformations

Common interview topics

  • What is LookML and what problem does Looker's semantic layer solve
  • Explain the difference between a dimension and a measure in Looker
  • How do you model a many-to-many relationship in LookML
  • What is a persistent derived table (PDT) and when would you use one
  • How would you onboard a business analyst to self-serve in Looker

Browse Looker jobs