Multidimensionality
An example makes things clearer. Imagine you start your own retail computer business. You decide to prepare a financial forecast or business plan showing your expected sales, costs, and expenses. You are planning on selling three types of equipment: laptops, tablets, and phones. One way to structure the data in a worksheet is to track accounts down the rows and the products across the columns.
| Phones | Tablets | Laptops | |
|---|---|---|---|
| Sales | 10,000 | 20,000 | 30,000 |
| Cost of goods | 6,500 | 13,000 | 19,500 |
| Gross margin | 3,500 | 7,000 | 10,500 |
As time passes, you want to calculate total product figures. You want to see the actual numbers next to the forecast numbers. Then you think how nice it would be to see this information month-by-month, and year-to-date, as well. You also open up five more stores and want to see the same figures displayed for each store and the total for all your stores. This is an example of multidimensional data.
A dimension in a multidimensional data set is a group of similar items that can be displayed as row or column headers in a report. With the above data, you can display Products across Time or Accounts across Time or Accounts by Store or Products by Store.
Read through the example again. There are five sets of similar items that can be displayed as row or column headings:
- Products
- Stores
- Time (months and years)
- Accounts (sales, costs of goods, gross margin and so on)
- Version (actual, budget, forecast)
These are the five dimensions of the model.
Multidimensionality provides the ability to rearrange these dimensions. You can choose which dimensions to display as column headings and which as row headings. With a multidimensional database such as OLAP, any of the above views of the data is available at the click of a button.