Intellicus Enterprise Reporting and Business Insights 18.1

Introduction to Analytical Objects

0 views December 28, 2020 0

Intellicus’ Analytical Object is a definition of cube comprising of related dimensions and measure groups used to analyze data.

Intellicus cubes are multi-dimensional, pre-aggregated data of your source data. The cubes are persisted in file system under Intellicus folders and are editable by data administrator.

To design and save an Analytical Object in a specific category, you need Read, Write and Execute permissions for Analytical Objects in that category.

This document explains the steps for a designer to design and build cubes using Analytical Object.



Ensure that Intellicus system you are using is licensed for High Speed View feature. This is an add-on feature in professional and enterprise editions of Intellicus.

Disk Space

Each cube can take from few KBs of disk space to many GBs based on certain factors used to design the cube. We will look at those factors later in this document.

Ensure you have enough disk space for all the cubes you plan to create and build.


Cube building is an activity of pre-aggregation. It requires sorting and other data processing activities. Intellicus uses minimum required memory at a given point in time, using streaming-in data and disk swapping.

Ensure that each cube being built in parallel has at least 1 GB of RAM available to it during the build time. This is considering building cube of 2-3 GB size from a 5-10 million transaction rows.

About Cubes

A cube is a structured multidimensional data-set; it has business dimensions and pre-calculates aggregations ahead of time for querying.

The structure of a cube makes it easy to visualize or conceptualize data along various dimensions of a cube making it easy to query and interact with the cube. Cubes organize data in a hierarchical arrangement, according to dimensions and measures.

Dimensions group the data along natural categories and consist of one or more levels. Each level represents a different group within the same dimension. For example, a time dimension can include levels such as years, months, and days.

Measures are the cube data values that are summarized and analyzed. A measure is the combination of a numeric input column with a roll-up rule or statistic. Attributes represent a single type of information in a dimension. For example, year is an attribute in the time dimension.

The elements of a dimension can be organized as a hierarchy—a set of parent-child relationships, where a parent member summarizes its children.

Let’s apply this to an example. For example, from your data warehouse you can create a cube which indexes and pre-computes sales data.

In your cube you could have all those pre-computed dimensions: sales by months, by week, by salesman, by client, by geographical region, by product model, etc. Then you can run queries on your cube to have the total, average and maximum sales by (month, salesman, region), or by (product model, region), or by (salesman, month). Since all the data is pre-computed and indexed, the queries are fast.

To design a cube, you can source both Fact Data (Measures) and Dimension Data from Intellicus Query Objects. Query Objects can fetch data from RDBMS or file sources. Refer WorkingwithQueryObjects.pdf for more details.

You can thereafter browse a cube on High Speed View or view High Speed reports that draws data from a cube.