Data Concepts Wiki
Short, fact-based reference entries on the BI and analytics concepts we mention throughout our BI Tools guides. Each entry covers one concept, links to related terms, and notes how Holistics approaches it where relevant.
Semantic Layer
How business meaning is defined, governed, and consumed across your analytics stack.
Analytics Query Language (AQL)
A metric-centric query language that operates at a higher abstraction level than SQL, designed to express complex analytical questions within a governed semantic layer.
Business Glossary
A shared vocabulary of business terms, their definitions, owners, and relationships – the organizational agreement on what words mean when applied to data.
Composable Metrics
Metrics that can reference other metrics, allowing complex calculations to be built from simpler governed building blocks rather than rewritten from scratch.
Conditional Governance
When a BI tool's AI is governed for simple queries but silently falls back to ungoverned text-to-SQL for complex ones, creating a false sense of security.
Data Contract
A formal agreement between a data producer and its consumers specifying the schema, quality guarantees, update frequency, and ownership of a data asset.
Governed Surface Area
The percentage of an organization's analytical questions that a semantic layer can answer natively, without workarounds like derived tables or custom SQL.
Headless Semantic Layer
A semantic layer that exposes metric definitions through APIs without requiring a specific BI front-end, letting any application query governed metrics.
Metrics Layer
A centralized system for defining, managing, and serving business metrics consistently across every analytics tool, notebook, and AI interface in the stack.
Push-down Semantic Layer
A semantic layer architecture where metric computation is pushed down to the data warehouse rather than executed in an intermediate engine.
Semantic Gap
The mismatch between the real-world meaning of data and its technical representation in databases, causing misinterpretation when humans or AI query it.
Semantic Layer Computation Limits
The specific analytical operations that a semantic layer cannot express natively, forcing logic into derived tables or workarounds.
Semantic Layer as Code
Managing semantic layer definitions in version-controlled code files rather than GUI configurations, enabling Git workflows, code review, and CI/CD for metric definitions.
Semantic Leakage
When business logic that should live in the governed semantic layer escapes into upstream transformations, downstream formulas, or analyst memory – causing metric definitions to fork and trust to erode.
The Semantic Ceiling
The point where a semantic layer stops answering analytical questions natively and forces logic into workarounds – derived tables, table calculations, or analyst tickets.
Universal Semantic Layer
A semantic layer that serves metric definitions across every tool in the data stack, not just the BI tool where they were defined.
What Is a Semantic Layer?
A centralized definition of business metrics and dimensions that sits between the data warehouse and every downstream consumer – dashboards, AI, APIs, and embedded analytics.
Self-Service BI
Patterns and capabilities that let non-engineers find, trust, and shape data independently.
Ad-hoc Reporting
The ability for users to create one-off reports and explore data on demand, without pre-built dashboards or analyst assistance.
Data Alerting
Automated notifications triggered when a metric crosses a defined threshold, delivering the right data to the right person without requiring them to check a dashboard.
Data Catalog
A searchable inventory of an organization's data assets – datasets, metrics, dashboards – with metadata about ownership, freshness, lineage, and usage.
Data Wrangling
The process of cleaning, structuring, and transforming raw data into a usable format for analysis, often the most time-consuming step in any analytics workflow.
Parameterized Report
A report template with dynamic inputs – date ranges, filters, segments – that let a single report definition serve multiple contexts without duplication.
Time-to-Insight Scorecard
A maturity framework measuring how quickly an organization can move from a question to a trusted, data-backed answer across four stages: discover, prep, build, and operationalize.
Embedded Analytics
Architecture and methods for shipping analytics inside your product.
Dynamic Database Routing
An embedded analytics pattern where queries are automatically routed to different databases based on the authenticated user's tenant, enabling physical data isolation.
Dynamic Schema Routing
An embedded analytics pattern where queries route to different database schemas based on the authenticated user's tenant, using a shared database with schema-level isolation.
Embed Events
Callbacks and event hooks that allow the host application to respond to user interactions within embedded analytics components, like filter changes, clicks, or drill-downs.
Embed Portal
A self-contained analytics portal embedded inside a host application, giving end users access to multiple dashboards, navigation, and exploration within a branded interface.
Embed Workers
Dedicated compute processes that handle concurrent queries from embedded analytics users, ensuring performance isolation between the host application's analytics load and internal BI usage.
Embedded Analytics Integration Methods
The technical approaches for integrating analytics into a host application: iframe, JavaScript SDK, API, and web components, each with different trade-offs in control, speed, and complexity.
JWT Embedding
A token-based authentication method for embedded analytics where the host application generates a signed JSON Web Token containing user identity and permissions.
Multi-Tenant Analytics
An analytics architecture where a single platform serves multiple customers or organizational units, each seeing only their own data through tenant-aware access controls.
Row-Level Security for Embedding
Data-layer access controls that restrict which rows each embedded analytics user can see, enforced at query time regardless of the dashboard configuration.
User Attributes for Embedding
Key-value pairs attached to an embedded analytics user that control data access, filtering, and personalization without modifying the dashboard definition.
White-Label Analytics
Embedded analytics that carries the host application's branding rather than the BI vendor's, making the analytics experience feel like a native product feature.