Dimensional models

The dimensional model is the part of the monitor model that contains the cubes that are used for dimensional analysis. The cubes in turn contain measures and dimensions. If you are planning to use Cognos for dimensional analysis in the dashboards, define a dimensional model.

You can generate multidimensional charts and graphs based on historical business performance data. In dimensional analysis, a large table of facts is described using measures and dimensions. Measures take on many values and participate in calculations; dimensions are entry points for manipulating and analyzing the data in meaningful ways.

As an example of dimensional analysis, consider an insurance company that has a claim process and uses Business Monitor to monitor each of the claims. Each claim is represented by a monitoring context instance and has metrics such as claim number, policy number, amount, type, status, claim date, country, city, and processing time. The company wants to analyze the claims at an aggregate level and visualize the data by various groupings. They use dimensional analysis to do so. They create measures, which include average claim amount, total claim amount, total number of claims, and average claim processing duration. These measures can be analyzed by specific dimensions. For example, they can look at the total number of claims of type auto in 2006, and can drill down to see the number of claims by month and day. They can look at the total number of claims that have not been paid. They can further analyze the data by country and city to get a better understanding of the nature of the claims and how the claim process is performing in various regions.