Looker Business Analyst key terms

Looker's modeling language for describing tables, columns, relationships, and reusable logic. Analysts write LookML so that everyone exploring data uses the same governed definitions instead of rebuilding calculations by hand.

A version-controlled collection of LookML files that defines models, views, and Explores for a Looker instance. Projects are typically synced from a Git repository so changes are reviewed and tracked.

A LookML file that maps a database table or derived table into fields such as dimensions and measures. Views are the building blocks that Explores join together.

A LookML file that declares database connections, includes view files, and defines Explores. It sets the boundaries of what users can query and how tables relate.

The user-facing starting point for a query, built from a base view plus joined views and defined in a model. Users pick fields from an Explore to build reports without writing SQL.

A LookML field that describes an attribute used for grouping or filtering, such as a date, category, or ID. Dimensions appear as groupable columns in an Explore.

A LookML field that aggregates rows, such as a count, sum, or average. Measures are computed by the database and appear as values rather than grouping options.

A LookML block that generates multiple date or time dimensions from one column, such as date, week, month, and year. It lets users drill into time periods consistently.

A LookML relationship between a base view and another view, declared with a SQL join type and relationship such as many-to-one. Joins determine which fields can be combined in one Explore.

A LookML setting on a join that declares cardinality, such as one-to-many or many-to-one. Correct relationships prevent duplicated rows and let Looker aggregate safely.

Row duplication caused by joining a one-to-many relationship into a query, which inflates sums and counts. Looker warns about fanout and offers symmetric aggregates to correct it.

A Looker technique that computes accurate aggregates across joined tables by wrapping values in unique keys. It prevents fanout from distorting sums and counts in multi-join queries.

The LookML declaration marking a field as the unique identifier for a view's rows. Looker uses it to detect fanout and to build symmetric aggregate logic.

A LookML view whose underlying data comes from a SQL query rather than a physical table. It supports persistent derived tables that are built and cached on a schedule.

A derived table whose results are written to a scratch schema and refreshed on a datagroup or trigger. It improves performance by avoiding repeated expensive subqueries.

A LookML-defined caching policy that controls when Looker invalidates cached results, often tied to an ETL schedule. Datagroups keep dashboards fresh without querying the warehouse constantly.

Looker's practice of storing query results for a set period so repeat requests return quickly. Cache settings balance freshness against warehouse cost and query load.

The panel in an Explore where users select dimensions, measures, and filters. It groups fields by view and supports searching, so users can find the right governed field.

A condition applied to a field that limits which rows a query returns, such as a date range or category list. Filters can be set in Explores, Looks, and dashboards.

LookML logic that customizes how a filter behaves, such as offering predefined time ranges or mapping a user's choice to SQL. It standardizes filtering for report consumers.

A LookML field that lets users choose an option which changes query logic, such as switching a measure or filtering a derived table. Parameters add interactivity without new Explores.

A templating syntax usable in LookML to insert values, such as filter selections or user attributes, into SQL or labels. It enables dynamic logic inside views and derived tables.

A saved query with defined fields, filters, and visualization that can be shared or added to dashboards. Looks are the simplest reusable reporting object in Looker.

A collection of tiles that display saved Looks, queries, or text in one layout. Dashboards support filters, scheduling, and sharing for at-a-glance reporting.

A single panel on a dashboard that renders a visualization, query, or text block. Tiles can be edited, resized, and linked to dashboard filters.

A control on a dashboard that applies one selection to multiple tiles at once, such as a date range or region. It keeps related charts consistent.

The chart or table type chosen to display query results, such as bar, line, or pie. Visualization settings control axes, series, legends, and formatting.

A LookML setting that defines where a user can click to see more detail, such as from a month to individual rows. Drilling supports investigation without new queries being built manually.

A configured link from one Explore to another, carrying filter values so users can jump to related detail. Drill paths connect summary metrics to underlying records.

A Looker feature that combines results from separate queries side by side, matching on shared keys. It is useful when metrics come from Explores that cannot be joined directly.

A calculation performed on query results after the database returns them, such as percent of total or running totals. Table calculations are written with Looker expressions.

The function syntax used in table calculations, custom fields, and filters, covering math, string, and date operations. It runs on results rather than in the warehouse SQL.

A measure a user creates inside an Explore or Look without editing LookML, often by filtering or combining existing measures. It supports ad hoc analysis while governed fields remain the base.

A request Looker sends to the database based on selected fields, filters, joins, and sorts. Understanding the generated SQL helps analysts debug results and performance.

A Looker interface for writing and running raw SQL against a connection, plus inspecting table schemas. Analysts use it to validate data and prototype derived tables.

A configured link between Looker and a data warehouse, including credentials, dialect, and default settings. Models reference connections to know where queries run.

A Looker capability that rewrites queries to use pre-aggregated rollup tables when possible, reducing scan size. It requires aggregate tables whose grain matches the request.

A value stored on a user or group, such as region or department, that LookML can reference to filter data automatically. It supports row-level security and personalized content.

A LookML field that restricts which rows a user can see based on their user attributes. It enforces row-level security inside Explores and dashboards.

A setting that grants a role the ability to perform actions such as exploring, editing LookML, or managing users. Permissions are assigned through roles and groups.

A named set of permissions and model access assigned to users or groups. Roles determine what a person can see and do in Looker.

A collection of users managed together for permissions, content access, and scheduling. Groups simplify administration compared with assigning rights user by user.

A container that organizes Looks and dashboards and controls who can view or edit them. Folder access works alongside roles to govern content sharing.

A Looker feature that checks saved Looks and dashboards for broken fields, missing references, or errors after model changes. It helps teams fix content before users hit failures.

A configuration that sends a dashboard or Look by email or to a destination on a recurring basis. Schedules depend on permissions and caching behavior.

Delivering Looker content inside another application using signed URLs or an embed SDK. Embedded users see governed data through restricted sessions.

Looker's programmatic interface for running queries, managing content, and administering users. Analysts use it to automate reporting and integrate Looker with other tools.