Previewer is a Seward Consulting Ltd. product which produces demonstration data for presentation in POWERPLAY, a Cognos business intelligence tool.

POWERPLAY is a Business Intelligence (BI) product which uses cubes to organize data about a business, normally extracted from or resident in a relational database, for analysis and presentation.

The POWERPLAY cube is a simple model of data organization. Those fields in a database whose contents appear in a cube are divided into two classes, dimensions and measures. A dimension is a field like "District Office", "Month sold", "Salesperson", or "Product type". It serves to label the record in which it is found. The possible values in that field are known as categories. A measure is a field like "Total sales", "Expenditures", or "Units sold". It contains a numeric quantity, and one for which totals and subtotals are meaningful.

In addition to actual values of the field specified as a dimension, other values are also accepted as categories. The categories for a dimension form a tree, and only the lowest-level nodes of that tree are the actual field values. Also, each category has a full name as well as an internal code. Thus, a cube contains summarized information that would be stored in multiple tables in a normalized production database.

With Previewer a user can quickly and easily define what a cube should look like in terms of dimensions, measures and overall percentages at each level in the dimension hierarchy. Previewer uses the resulting model to generate data points which can then be converted to POWERPLAY cubes. Cubes created by Previewer are generated randomly but constrained to the specified percentages. In most cases, Previewer cubes are virtually indistinguishable from cubes based on actual production data.

Whereas someone uses POWERPLAY to enhance their knowledge of a business, someone can use Previewer and his knowledge of the business to create fabricated POWERPLAY cubes that look realistic.

The principle behind Previewer is that you set up the structure of each dimension and measure and then you enter the relative weightings at the various levels in the dimension hierarchy. This weighting information, called "percentages", is used by Previewer in the actual generation of cubes. By specifying the percentages you can influence how the cube will be generated. The only rule is that at each level the total weighting has to add up to 100%. For example, the years dimension, 100%, might be composed of the following: 1998 (25%), 1999 (25%), and 2000 (50%). You could further split each year into quarters with their own respective percentages reflecting the overall trends of the business model.

You also specify for each measure a total value. For example, the Sales measure might be $2,545,000 and the Cost measure might be $1,914,000. Previewer uses these amounts and randomly generates points that are constrained by the percentage values associated with each level in the hierarchy.

Previewer can be used to demonstrate the operation of Business Intelligence technology on data customized to an individual customer's operation without the need for the lengthy preparations to connect to actual live databases. This provides pre-sales reps of POWERPLAY products the ability to quickly show the technology based solely on prior knowledge of the business.

Another use of Previewer is in creating industry vertical demos. POWERPLAY portals and sales kits based on Previewer cubes or star-schemas can be constructed efficiently and without the problems connected with the sourcing of proprietary data. Previewer also has the ability to reverse engineer cubes so that existing POWERPLAY data can be leveraged into the sales cycle.

In many situations Previewer can also be used by end-users of POWERPLAY to prototype new cubes or adjust existing cubes before resources are expended in applying the changes in the source data mart or data warehouse.

The following are major features currently support by Previewer 2.0:

• Intuitive user interface.
• Maintenance of Dimension and Measure hierarchies.
• Percentage bars for specifying relative differences between Categories.
• Reverse engineering of existing cubes.
• Measures that can be defined in terms of other measures.
• Importing of dimensions from other cubes.
• Facility for creating star schemas in any ODBC database.
• Control of irregularity at the cube and dimension levels.

The following is a screen snapshot of the Previewer interface: