In this post we now create our first dimension. A dimension is the data on which the question is being asked. The purpose of the cube is to answer questions hence the terms Business intelligence or Decision Support. A typical question that might be asked of a cube is “What are the top selling products for the last 6 months?” Here the question being asked is about Product and dates what products and what dates? These are your dimensions. The next part of the question is the quantifier. Here is having a number of quantifier that apply to sales such as profits, unit’s price or sold quantity. In this case, of the 3 measureable attributes listed we are interested in the measure Sold Quantity since we need to know the top selling items. So let’s look at the dimension called product now. As you can see there are 2 different dimensions here Products and Dates, the idea being that we first need to display the dates for the last 6 months and then within those dates identify the top selling products The order in which the dimension are placed will impact the results but for now let’s simply look at the dimension as it is.
The product dimension contains attributes that are related to the products. A product will have a number of properties associated with it such as Unit price, Colour, description, packaging information etc. All properties which can be considered as Meta data that defines the dimension will be part of the dimension e.g. email address is not a property of a product , a sofa cannot have an email id. So in order to defined our dimension the first thing we need to ask ourselves is
What is this dimension about? Do you need this dimension within the cube? How will the dimension be used? Let’s answer these questions now.
The product dimension is a dimension that will contain all information related to the products sold by the company. It will contain information such as colour, description, product classification, if it is currently being manufactured and how long with will take to ship the product.
Yes, the dimension is needed within the cube in order to analysis sales information by product and identify products that need to be restocked etc.
These dimensions will be used in a number of reports where the business user might need to drill down to get further product level information and summary products by sales or any other measure as required. The product catalogue can be used by a variety of systems and therefore needs to be complete and accurate and should store all possible states of the products including previously sold products that are now defunct or products for whom the pricing or some other information has been modified. This information is captured as a slowly changing dimension which will be explained later.