Skip to main content

Looker Guide

Looker transforms how organizations use data by making it more accessible and actionable for teams across the business. This guide explains how Looker creates a single source of truth while enabling self-service analytics.

Looker Architecture​

Looker Platform Architecture

Core components of the Looker platform and how they interact

100%
🔍 Use Ctrl+Scroll to zoom
Connects toDefinesPowersEnablesSupportsManagesDatabaseConnectionsLookMLModelsExploresDashboardsEmbeddedAnalyticsAPIServices

Legend

Components
Data
Model
Interface
Visualization
Integration
Connection Types
Connection
Definition

What Makes Looker Different​

📊

LookML Data Modeling

Technical

LookML creates a semantic layer that translates complex database structures into business-friendly terminology, providing:

  • A single source of truth for business metrics across the organization
  • Consistent definitions that prevent conflicting reports
  • Empowerment of non-technical users to explore data safely
  • Reduced reporting bottlenecks from centralized BI teams
📊

LookML Data Modeling

Non-Technical

LookML is a language for describing dimensions, aggregates, calculations, and data relationships in a SQL database. Key technical aspects include:

  • Git-integrated version control for model development
  • Reusable model components via extends and refinements
  • Centralized business logic with derived tables and SQL Runner
  • Advanced SQL generation with symmetric aggregates and fan-out control
🔌

Embedded Analytics

Technical

Embedded analytics transforms applications by bringing data directly to users in context:

  • Increases user engagement through contextual insights
  • Creates new revenue streams through data monetization
  • Reduces development effort for in-app analytics features
  • Maintains consistent metrics between internal and customer-facing analytics
  • Enables "analytics as a product" business models
🔌

Embedded Analytics

Non-Technical

Looker provides comprehensive embedding capabilities for integrating analytics into applications:

  • iframe embedding with signed URLs and SSO authentication
  • JavaScript API for two-way communication with embedded content
  • Components SDK for embedding specific visualization elements
  • Extension Framework for custom applications within Looker
  • White-labeled embedding with custom themes and branding

LookML Development Guide​

Key LookML Concepts​

LookML models include these key components:

  • Connection: Defines the database connection
  • Model: Container for explores and views
  • View: Defines a table or derived table
  • Dimension: A field that represents an attribute
  • Measure: An aggregate calculation (sum, count, average)
  • Explore: Joins multiple views for querying

Example LookML Model​

connection: "my_database"

include: "/views/*.view.lkml"

# Base model for e-commerce analytics
model: ecommerce {
  
  # Define a view for the orders table
  view: orders {
    sql_table_name: public.orders ;;
    
    dimension: order_id {
      primary_key: yes
      type: number
      sql: ${TABLE}.id ;;
    }
    
    dimension_group: created {
      type: time
      timeframes: [date, week, month, year]
      sql: ${TABLE}.created_at ;;
    }
    
    measure: count {
      type: count
      drill_fields: [order_id, created_date]
    }
    
    measure: total_revenue {
      type: sum
      sql: ${TABLE}.amount ;;
      value_format_name: usd
    }
  }
  
  # Define the explore that users will query
  explore: order_analysis {
    from: orders
    
    join: users {
      type: left_outer
      sql_on: ${orders.user_id} = ${users.id} ;;
      relationship: many_to_one
    }
  }
}

Looker Administration​

🔒

Looker Security Model

Technical

Looker's security model balances accessibility with governance:

  • Ensures sensitive data is only visible to authorized users
  • Provides customized views based on user roles or departments
  • Maintains regulatory compliance with GDPR, HIPAA, and other frameworks
  • Reduces risk of data leakage through consistent policy application
  • Enables safe self-service without compromising security
🔒

Looker Security Model

Non-Technical

Looker security implementation details:

  • Role-based access control with user attributes
  • Row-level security via access grants and datagroups
  • Content access controls with folders and permissions
  • SAML/OIDC integration for SSO authentication
  • API key rotation and OAuth management
  • Field-level permissions via LookML access_filters

Looker Performance Optimization​

Query Performance​

Optimize Looker query performance with these techniques:

  1. Database-Level Optimizations:

    • Create appropriate indexes on join and filter fields
    • Use materialized views for complex calculations
    • Implement proper partitioning strategies
    • Use distribution keys in cloud warehouses

  2. LookML-Level Optimizations:

    • Create aggregate awareness with PDTs
    • Implement datagroups for cache control
    • Use symmetric aggregates to prevent fan-out
    • Deploy clustering keys in persistent derived tables

  3. Front-End Optimizations:

    • Limit the number of dimensions and visualizations per dashboard
    • Use dashboard filters effectively
    • Implement dashboard level caching policies
    • Configure appropriate query timeout settings

Enterprise Deployment Best Practices​

Looker Enterprise Deployment Architecture

Recommended deployment architecture for enterprise Looker implementations

100%
🔍 Use Ctrl+Scroll to zoom
CommitTriggerDeployPromoteConnectConnectConnectProductionInstanceStagingInstanceDevelopmentInstanceGitRepositoryCI/CDPipelineDataWarehouse

Legend

Components
Deploy
Tool
Data
Connection Types
Deployment Flow
Data Connection

Development Workflow​

For enterprise implementations, follow this Looker development workflow:

  1. Development Instance:

    • Developers create and modify LookML models
    • Unit testing with SQL Runner and Explores
    • Content validation with test users

  2. Version Control Integration:

    • Commit changes to Git repository
    • Pull request workflows for code review
    • Branch management for feature development

  3. Staging Environment:

    • Automated testing of deployment changes
    • Performance testing with production-like data volumes
    • UAT with selected business users

  4. Production Deployment:

    • Scheduled deployment windows
    • Rollback procedures for failed deployments
    • Post-deployment validation

Additional Resources​