Monitor and Cognos
IBM Business Monitor supplies a copy of IBM Cognos as part of the product solution. Cognos is a Business Intelligence and Analytics engine. From a Monitor perspective, Cognos provides the capability to define and view dimensional models.
It is the Business Space Report Viewer widget that displays the model.
Cognos has a bewildering number of components and tools. Many of these don't identify themselves so you simply have to know which one is which. For example, when looking at the Report Viewer widget, the Edit Report menu item:
Will bring up a web based tool called IBM Cognos Business Insight Advanced but there is nothing on the page to let you know that this is what you are in. (Later: Actually, I take that back … in the Help menu there is an about box).
Cubes, Categories, Dimensions and Levels
Experience has shown that the following topic can be difficult for many folks to understand. Let us examine a scenario. Imagine a sales record that may contain fields as follows:
US State of sale
City of sale
Amount of sale
Category of sale
Now imagine that over time instances of such records will arrive. After, say, 6 months worth of data, we now want to perform some analysis of what we have found. What kinds of questions can we answer? Here are some possibilities:
Answering such questions from the examination of this data is where Cognos comes into play. In order to also start answering these questions we need to delve into cubes, measures and dimensions.
Lets start with measures. Measures can be thought of as a roll-up of a metric such as sales amount. This takes all the instance data values of the metric and has a function applied to them. The functions available include:
Measures are defined in the Dimensional Model tab of the Monitor Model editor. We define the name of the measure to be created and which metric is used for its creation along with the aggregation function to be applied to those metrics.
When charting, think of the measure as the value of the Y axis.
If measures are aggregates of metrics, dimensions can be thought of as the selections of which measures to aggregate together. For example, if we have a measure that is the sum of sales amounts and we have a dimension which is the US state in which the sale occurred, then we can determine the value of all sales by state.
When we define a dimension, we also define levels. The Levels define the levels of drill down. Every dimension must have at least one level. For example, if we imagine a dimension called "location" used to define the location in which a sale occurred, we might define a level as being the "US State". We will then be able to see all sales by US State. If we define a second level which is "City", then we will be able to drill down each US State to the cities from that state which have records.
There are an amazing number of chart types available in Cognos. To see a summary list, select the Insert Chart icon and have a look:
Here are some key terms to help us understand charting.
The Data Series can be thought of (loosely) as the X-Axis values in a bar chart. They are the entries that have specific values. For example, a data series for sales may be 2009, 2010, 2011 where each element in the series has a value (for examples, sales in that year). Each element in the Data Series has its own color or distinguishing characteristic.
The primary axis can be though of as the Y-Axis. This is where values will be plotted against the elements in the Data Series. It may help to think of the X-Axis as holding qualitiative data (years, regions, etc) while the Y-Axis holds quantitative data (sales, number of customers, etc).
The Legend is the key to the patterns or colors used to indicate the Data Series. The following image illustrates a Legend.