Pyramid Analytics
Glossary

PYRANA engine

Pyramid's query, analytics and calculation engine, pronounced like the fish. It builds and runs queries and calculations directly against your data sources, generating native SQL or MDX as needed, which is what lets Pyramid work at speed on large data without copying it out.

PQL

Pyramid Query Language, built into PYRANA. Structurally like MDX but optimised to run as SQL on whatever database sits underneath, and it carries the libraries for modelling, dynamic text and logic.

MDX

A long-standing language for querying multidimensional data and OLAP cubes. Pyramid uses MDX natively against cube sources, and PQL for everything else.

In-memory database

Pyramid's own in-memory engine, for when sub-second response is needed. You can model data into it, or leave data in the source and query it directly.

Direct query

Querying a source where it lives, without copying anything into Pyramid first. PYRANA pushes the logic down to the source, which avoids duplicating data and keeps results current.

Query sharding

Splitting a large query into parts that run in parallel, one of the ways PYRANA keeps performance up on big datasets.

Mini-services architecture

Pyramid runs as a set of separate services that can scale up and out independently, so the platform adapts as users and workloads change.

Web Server

The service that powers the browser interface. It does little processing and can be load-balanced across several machines.

Runtime Engine

The core server-side service that handles everything except batched work. Most interactive activity runs through it.

Task Engine

The service that handles batched work: printing, data preparation jobs and publications.

Router

The service that directs traffic between the others. At least one is required, and two are advised for fail-safe.

DS/ML Server

The service that runs all data science and machine learning, including Python and R engines.

NLP Server

The service that runs natural language querying and the other language-driven AI features.

GIS Server

The service that drives geospatial analysis and maps.

Solve Server

The service that runs the Solve optimisation engine.

Pulse

Pyramid's real-time engine, which pushes notifications when data crosses a condition you set.

Repository database

The database Pyramid uses to store its own content, settings and usage data, held in PostgreSQL, Oracle or SQL Server.

Multi-server deployment

Spreading Pyramid's services across several machines for performance, load balancing and failover, as opposed to a single-server install.

Failover and load balancing

Running more than one of each service so work is shared and the platform keeps going if a server fails.

Kubernetes and elastic scaling

Running Pyramid as containers that auto-scale up and down with demand, for predictable performance under changing load.

REST API and embedding

A programming interface for automating Pyramid and embedding its analytics inside other applications.

Deployment options

Pyramid runs self-hosted on premises, in the cloud on AWS or Azure, or hybrid, and is device and operating system agnostic.

Model

The module where you prepare data and build data models with no code. Pyramid's equivalent of the modelling layer in Power BI.

Data Flow

Connect to one or more sources, clean and shape the tables, then either query them directly or load them into a model. Where data preparation happens, without code.

Data source

A connection to where data lives: relational databases, OLAP cubes, data lakes, files or unstructured sources.

Data model

A prepared, structured dataset reports are built on, either modelled into the in-memory engine or mapped as a direct-query model over the source.

Data mashup

Joining tables from several different sources into one model, so a report can draw on more than one system at once.

Target node

The end point of a Data Flow that defines where the prepared data lands, such as the in-memory database or a relational store.

Virtual semantic model

A modelled layer that hides the complexity of the source and presents clean hierarchies, measures and definitions. Pyramid's semantic layer.

OLAP cube

A multidimensional structure for fast slicing and aggregation. Pyramid can build its own and query existing cubes, such as Analysis Services or SAP BW, natively.

Dimension

A way of cutting your numbers, such as product, region or time. The 'by' in sales by region.

Hierarchy

A drill path within a dimension, such as year to quarter to month, that lets you move between summary and detail.

Attribute and element

An attribute is a property within a dimension; an element is a single value of it, such as one product. The things you select to build a query.

Measure and member

A measure is a number you analyse, such as revenue; a member is a single value within a dimension. The building blocks of any query.

Aggregation

How values roll up, such as sum, average or count, when you move from detail to summary.

Granularity

The level of detail a model holds, such as a row per transaction or per day. It sets the floor for how finely you can analyse.

Refresh and process

Updating a modelled, in-memory dataset with the latest data from the source, on a schedule or on demand.

Smart Modeling

An AI-assisted way to build a model, using heuristics to suggest relationships and cut the manual transformation and cleansing work.

Fill In the Blank

A feature that fills gaps and nulls in a dataset, useful for time series and trend analysis where missing points would break the maths.

Data ingestion

Copying selected tables into a Pyramid data model, usually the in-memory database, when you want a modelled, optimised store rather than querying the source directly.

Discover

The data discovery and visualisation module. Build queries, slice and drill, and explore. Several ways to drive it, from drag-and-drop to natural language.

Formulate

Where you build calculations, KPIs, member lists, sets and scripts, in PQL or MDX, saved for reuse across reports.

Illustrate

For data-driven infographics, text and images that react to the data, including fully custom visuals built with D3 JavaScript.

Present

For building interactive dashboards and slide-style presentations, or storyboards, from content made in the other modules.

Publish

For pixel-perfect, highly formatted documents in PDF, Excel and other formats, scheduled and distributed at scale.

Tabulate

A spreadsheet-style interface for business modelling and planning, without exporting to Excel.

Solve

For business modelling and optimisation, working through scenarios to a recommended answer.

Discovery

A single analysis or report built in Discover. The basic unit you build, save and share.

Drop zones

The areas where you place dimensions and measures to build a query, such as rows, columns and colour. The query takes shape as you fill them.

Elements tree

The browsable list of the individual values, or elements, within a hierarchy, used to pick exactly what goes into a query.

Slicer and bookmark

A slicer is an interactive filter; a bookmark saves a set of slicer selections to reuse.

Quick Calc

On-the-fly calculations built in Discover from your current selection, covering grouping, formulas and lists, without opening Formulate.

Semantic calculation

A calculation that becomes part of the query sent to the engine, as opposed to one worked out from what is on screen.

Context calculation

A calculation worked out from what is on screen, such as a running total or share of parent.

List

A saved, logical set of elements, such as a custom group of products, built once and reused.

Set

A defined collection of members used in calculations and queries, the building block behind many advanced analyses.

Parameter

A value chosen at runtime, such as a Top N or a date, injected into a query or formula to make it dynamic.

Custom member and measure

A calculation saved as a reusable element, written in PQL or MDX, available across reports, dashboards and publications.

Formulation

A saved custom formula, kept in the CMS and dragged into analyses wherever it is needed.

Quick Discovery

Opening a formula or parameter straight into a Discover query, added to the drop zones automatically. A fast path from logic to analysis.

Visualisation

A chart, grid, map or other visual within a discovery. Pyramid ships a wide library and supports custom ones.

Grid

A table visualisation, Pyramid's pivot-style view for showing numbers in rows and columns.

Conditional formatting

Rules that colour or style a visual based on its values, such as red for figures below target.

Interactions

The links between visuals on a dashboard, so selecting in one filters or drives the others.

Drill down and drill through

Drill down moves from summary to detail within a hierarchy; drill through jumps to the underlying rows behind a number.

Custom visualisation

A bespoke visual built with D3 when the standard library does not cover a need.

GenBI

Pyramid's generative AI: using plain language and LLMs to generate queries, calculations, spreadsheet formulas, designs and narratives across the platform.

NLQ chatbot

The natural language query chatbot. Ask in ordinary words and Pyramid finds the right hierarchies and measures and picks a visualisation. It runs on any source and uses LLMs.

Smart Insights

AI-generated, plain-language analysis of a visualisation, slide or whole model, explaining anomalies, drivers and relationships. It works on any SQL or MDX source.

Auto Discover

An AI feature that surfaces correlations, commonalities and anomalies automatically. A fast way to find what is worth a closer look.

Explain

A natural language feature that puts the story behind a result into readable text.

Data Science Workbench

Pyramid's environment for professional data scientists, supporting Python, R and machine-learning libraries alongside the no-code tools.

Python and R scripts

Scripts woven into models and analyses, so machine-learning logic runs inside the platform rather than in a separate tool.

Algorithms

Machine-learning steps you can add inside a Data Flow or analysis, such as clustering or regression, without leaving Pyramid.

Multi-LLM strategy

Binding different large language models to different data models - for example a domain-specific model for a specialist area - rather than relying on one.

Automated analysis

Built-in analytics Pyramid runs for you, including forecasting, Pareto, category and trend analysis, surfaced as you explore.

Content Management

Where Pyramid stores content and controls access, sharing, versioning and lineage. The backbone that keeps a shared deployment governed.

Publication

A formatted document or set of documents produced in Publish and distributed, on a schedule or on demand.

Report bursting

Producing one report template and sending each recipient their own slice of the data, personalised, from a single design. Pyramid's take on enterprise reporting.

Subscriptions and scheduling

Automated delivery of content on a schedule, by email or to a destination, so people get what they need without logging in.

Dashboard and storyboard

A dashboard is a single interactive view of several visuals; a storyboard is a sequence of them, built in Present for sharing or presenting.

Infographic

A data-driven graphic built in Illustrate, where shapes, text and images respond to the numbers.

Dynamic text

Narrative that updates with the data, so written commentary and headings stay in step with the numbers.

Alerts

Conditions you set so Pyramid notifies people when something in the data changes or crosses a threshold.

Roles

The heart of Pyramid's security. Users are placed in roles, and roles govern access to data, content and sharing across the platform.

Profiles

Settings that decide which modules and functions a Pro user can use, set by administrators per user or role.

Tenants and multi-tenancy

Separate, isolated environments for different groups or customers within one Pyramid platform. Users belong to one tenant unless cross-tenancy is enabled.

Domain Admin

An administrator with partial rights limited to a specific tenant, rather than the whole platform.

Workgroup and Public content

The shared content areas. Workgroup folders belong to roles; public folders allow per-role read, write and management rights for finer control.

Content security

Per-folder and per-item access set as read, write or management rights, with the option to propagate settings down a folder tree.

Data security

Rules controlling which rows of data each user can see, so one report can be shared widely and safely.

Authentication and single sign-on

How users prove who they are, supporting single sign-on and multi-factor authentication, separate from what they are then allowed to do.

Admin console

The central place administrators manage servers, users, roles, profiles, security and content.

Data Manager

The admin area for managing data sources, models and server connections, alongside the Content Manager for content.

Usage analytics

Built-in reporting on how the platform is used, drawn from the repository, so administrators can track adoption and performance.

User types

The four licence tiers - Professional, Analyst, Viewer and Basic - from full authoring down to view-only, which decide what each person can do.

Enterprise edition

Pyramid's top edition, which unlocks features such as natural language query. Edition and licence together set each user's capabilities.

Embedded edition

A way of licensing Pyramid to embed its analytics inside another product or portal, common for software vendors.