September 18, 2026 · Vistelio
Merged Results, cross-database joins, and modeled alternatives all promise to combine data from two Looker Explores or two warehouses. They behave very differently. This tutorial shows how each one actually executes, where the numbers go wrong, and a decision path for choosing between them.
Read moreSeptember 17, 2026 · Vistelio
A practical tutorial on building maps in Looker: the location, zipcode and region dimension types, distance measures, map_layer and custom TopoJSON layers, BigQuery GEOGRAPHY columns, and the performance and permission traps that make Looker maps slow or blank.
Read moreSeptember 15, 2026 · Vistelio
A practical tutorial on LookML localization: locale_data strings files, localization_settings, locale-aware labels and descriptions, translating LookML dashboards, driving locale from user attributes for embedded analytics, and the governance habits that keep translations from rotting.
Read moreSeptember 14, 2026
BI Engine is the cheapest dashboard-latency win available to a Looker team on BigQuery, and most instances use it badly or not at all. This tutorial covers how BI Engine acceleration actually works, which Looker query patterns it accepts, how to size a reservation, how to prove hit rates from INFORMATION_SCHEMA and System Activity, and how to model LookML so your tiles stay in the accelerated path.
Read moreSeptember 11, 2026 · Vistelio
LookML has Git. The dashboards, Looks, folders, boards and schedules your business actually opens do not. This tutorial shows how to export and import user-created Looker content between dev, QA and production instances with gzr and Looker Deployer, how to remap folders and models, and how to keep the whole thing in a repeatable pipeline.
Read moreSeptember 10, 2026 · Vistelio
Most Looker tuning advice assumes BigQuery. This tutorial walks through what changes when your LookML model sits on Snowflake or Databricks SQL: connection settings, temp schemas for PDTs, warehouse sizing, query killing, and the dialect-specific LookML you should be writing.
Read moreSeptember 9, 2026 · Vistelio
Every mature Looker project accumulates fields nobody should use any more: revenue_v2, the metric finance disowned, the dimension that was a typo in 2022. Deleting them breaks saved Looks, schedules and embedded tiles. This tutorial walks through a repeatable deprecation workflow for LookML: proving a field is unused, staging a rename with aliases, using the Content Validator to repair saved content, and closing the window with a clean removal.
Read moreSeptember 9, 2026 · Vistelio
A practical guide to LookML dashboards: when to use them instead of user-defined dashboards, dashboard file syntax and the listen block, converting existing UDDs with Get LookML, extends for regional variants, and using the Content Validator as a deployment gate.
Read moreAugust 26, 2026 · Vistelio
Model Context Protocol is how AI assistants get tools. This tutorial stands up a Looker MCP server, connects Gemini CLI or Claude Desktop, walks through the tools it exposes, and covers the credentials, permissions and LookML shaping that decide whether the answers are trustworthy.
Read moreAugust 25, 2026
Looker is a modern, cloud-based business intelligence (BI) platform designed to explore, share, and visualize data at scale. This platform leverages a data modeling language, LookML, to describe and interact with data, making it a compelling choice for companies looking to democratize data insights. This article provides a technical introduction to Looker, discussing its architecture, core components, and capabilities.
Read moreAugust 25, 2026
Standing up Looker in 2026: provisioning a Looker (Google Cloud core) instance from the Cloud console, Google OAuth sign-in, connecting BigQuery or another warehouse with a service account and PDT schema, creating a LookML project in Git, and the day-one settings that save pain later.
Read moreAugust 25, 2026 · Vistelio
How to use the System Activity explores (History, Field Usage, PDT Event Log, Database Performance) to find slow dashboards, prune unused fields and content, attribute warehouse cost to users and schedules, and feed the results back into the model.
Read more