Dimensions, Members, Cubes and Templates

These terms may be familiar if you have previously been involved in a Business Intelligence or “OLAP” project. OLAP stands for OnLine Analytical Processing – which is pretty unhelpful as a description but if you are aware of the terms “Member” and “Dimension” much of what follows should be familiar.

Here is a video, or read below.

Dimensions & Members

A dimension is a way of looking at your data, a perspective. We could analyse our data by year, month or customer. By salesman, geography or department. Each of these are different perspectives and thus different dimensions.

Dimensions are represented within FastClose using a D in a blue ball.

Dimensions are made up of members. For example a Month dimension would be made up of the members: January, February, March, April, May, June, July, August, September, October, November and December.

A year dimension whilst conceptually having a more or less infinite number of members would likely be limited to those that are relevant from a business perspective for example 2010 through 2025 say.

A customer dimension might contain a member for each of your customers. It is common for such a member to have what is called a code and a caption. For example “Kermit’s Lilly Pad Co.” might use that as the Caption but use an identifying code such as 00010, say, to reference it, which is shorter and easier to type.

Members are represented within FastClose using an M in a blue ball .

Cubes & Templates

Dimensions are arranged into cubes to organise data and enable users to answer a variety of business questions.

Unless you are a report template designer you are unlikely to work with cubes directly in FastClose but will access them through predefined templates, arranged into modular groups for areas like General Ledger (GL), Accounts Receivable (AR), Accounts Payable (AP) etc… which can be found in the “New Report” dialog. Templates are starting positions from which to build new reports.

Dimensions are reused and often appear in many different cubes.

For example practically every cube will have some sort of Year and Month dimensions whilst a customer dimension would appear in GL and AR cubes as well as elsewhere.

“Measures” – A Special Dimension

The members of a dimension remain constant despite appearing in many different cubes. The one exception to this is the “Measures” dimension. This is a special dimension that every cube has and is particular to each cube.

Measures are database columns that you can add up. So the measures dimension in a GL Balance cube will have quite different members to that in an Inventory cube because you are adding up different things. It is this feature that enables a GL Balances cube to answer questions about GL balances and an Inventory cube to answer questions about Inventory.

For example an Inventory cube might have a measures dimension that includes members such as “Quantity”, “Total Cost” and “Lead Time”, items that help answer questions about inventory whilst a GL Balances cube might have a measures dimension that includes the members “Amount” and “Budget” that would help a user understand their actuals in the GL Chart of Accounts against numbers previously budgeted. So you have different measures dimensions in different cubes helping answer different questions.

The “Measures” dimension is represented within FastClose using a $ in a green ball . The dollar sign is used because Measures are most commonly financial.

Show me the numbers!

As discussed, the “Measures” dimension specifies members that can be added up and analysed, be they actuals, budgets, quantities or whatever. These are the things we are measuring.

What we are measuring by, are the other dimensions such as company, division or region, or by year and period.

Combining the measures dimension with these other dimensions allows us to form a cube with which we can analyse and break down our data, combining numbers measured by one dimension to form new totals, whilst splitting by others to reveal detail. We will talk about this a lot more in the sessions that follow.

Advanced Topic: Attributes

Originally, OLAP focussed on the idea of Dimensions and Members to describe data. But after a few years it became clear that a further concept was required to provide additional information about individual members within a dimension, so the idea of the “attribute” was born.

It is very common for a dimension to have additional “attributes”. These are properties that relate, describe and belong to it. Attributes may not be textual, but most are and can be thought of as additional fields providing further information regarding a member of a dimension.

Attributes are defined at the dimension level, so if one member has an attribute all members of that dimension will have that attribute. So it isn’t the case that only a few customers might have “Name” attributes. If the “Name” attribute is defined on the customer dimension, then all members will have a name, and can fill that attribute in. This makes it possible to display or query them for every member of a dimension.

For example, a customer dimension’s attributes could include company name, their address, city and zip code. Eg:

As well as being useful in reports, displayed alongside in this case, the customer concerned, attributes can also be useful to select all members having a particular attribute value. So here we could use the Post Town, to look at beer provided to all the pubs in Rugby as a whole.

Summary

All the terms we have used so far: Members, Dimensions, “Measures”, Cubes and so on, are all things that describe the data and the ways that it can be analysed in FastClose. They describe a dataset and make it available as a data source in FastClose; a collection of numbers and perspectives (dimensions) with which to understand them. But, they tell us nothing about the layout and organisation of reports that may be built using these data sources – and that is where we will go in the next session.