Main PowerPlay Transformer features:

Advanced modelling capabilities Leverages aggregate and fact tables Intergation of the query data into the model The models are defined by dimension structures and measures which can be easily customized No custom programming required Time periods are handled in a specific way which makes data delivery easy High flexibilty and portability

PowerPlay cubes modeling and development process

The application development process in Cognos PowerPlay Transformer usually contains the following steps:

Gather business users requirements Create a Model Plan which will include design and data requirements for the model and decide on the cube distribution methodology Select measures Plan dimensions and levels Design and develop the PowerPlay model in Transformer Create PowerPlay cubes and distribute them to the end users Analyze and explore the cube data using PowerPlay

A Transformer Model can be also created using a wizard with AutoDesign function or manually. We strongly recommend to avoid the AutoDesign function as in most cases it does not provide the results expected.

PowerPlay Transformer application

The default Cognos PowerPlay transformer window is composed of the following windows depicted on the the screenshots below. Building a PowerPlay model involves designating columns in the source files as measures and dimensions&levels. Data Sources - a list of all connected data sources. Each source can be expanded and it lists all available columns. Users can browse the source data by using Data Source Viewer. Columns from a data source window can be dragged & dropped to the dimensions and measures windows accordingly.

 Dimension Map manages the dimensions and levels hierarchy in the data model. It allows creating new dimensions and modifying existing using an user-friendly graphical interface. Dimensions and underlaying Levels are created by moving columns from the Data Sources window to the corresponding dimension column. Measures - drag & drop columns to manage number figures (facts) in a PowerPlay model. PowerCubes - you can specify here a list of cubes created by the model. User Classes - define the security. Signons - datasource connection information. Cognos Powerplay Transformer model for the palm nursery business scenario:

Cognos PowerPlay Transformer Guide - data sources

PowerPlay Transformer is supplied with the source data through the data definitions managed in a Data Sources window. Source data typically comes from a data warehouse, sometimes from production ERP system and in many cases includes multiple sources in order to deliver information required by the model.

Data sources recommendations

Column names in a data source must be unique. If a model contains multiple data sources, then the columns can be duplicated. Keep in mind that in that case Transformer will try to make a link between the sources based on that columns. Every dimension requires a column on a lowest level of the dimensions hierarchy (usually an ID column is at the bottom) At least one measure must be derived directly from a source column. Other measures can be calculated. A good practice is to keep column names in the model consistent with the data source column names. Another good practice is to preview source data with the Data source viewer. Avoid changing or switching columns in the source data PowerPlay Transformer uses the following data types to populate dimensions, levels, categories and measures: Date - a special type which defines time periods and can be linked directly to the time dimension which is automatically managed by Transformer. Usually it is a native database date format, sometimes text in YYYYMMDD format. Numeric - used mainly by measures Text - defines levels and dimensions

Cognos Powerplay Transformer - multiple data sources in a model with the data source definition:

The data source properties window also allows users to choose the datasource type, point to the right file or database server and set up a number of processing options, like: define timing, uniqueness verification, specify number separators and indicate current period in the time dimension. We will get into more details on that options later on in this tutorial.

Supported Data Sources by Cognos Transformer

Impromptu query definition files (.iqd) - which can query local or server-based databases

 Relational sources - including Oracle, SQL, IBM, Teradata, Sybase, and any other ODBC source Dimensional sources - Cognos OLAP, SAP BW, Microsoft SQL Server Analysis Services, Essbase, Oracle 10G, IBM DB2 CubeViews ERP systems - SAP, PeopleSoft and Siebel XML, Java beans, JDBC, LDAP, WSDL Excel files, Access files and csv or text files Legacy and Mainframe systems - like VSAM, IMS, IDMS, Cobol Copybooks Content management data - FileNet, Documentum, and OpenSoft

Time dimension
Time dimension is a special type of dimension. Cognos provides an automatic support for handling time dimensions which is working very well.

Go to Tools -> Date wizard (or use a date wizard icon located in the toolbar) and follow the wizard steps (values entered in our example in blue): 1. Provide name of the dimension - TIME 2. Choose Date Source Type ( Column ) and point the coumn name which contains date values ( DT ). 3. Do you want years in the dimension? Yes 4. Choose the kind of years between Calendar years (355 or 356 days) and Lunar years of 52 weeks 5. Do you want quarter-years in the time dimension? Yes 7. Do you want months in the time dimension? Yes 9. Do you want months in the time dimension? Yes 10. Do you want days in the time dimension? No 11. Choose the first day of the week. Default is Sunday, we will choose Monday 12. How to treat a week which spans a month? Split between the two 13. Enter ther first day of a particular year (yyyymmdd) - year value in this setting is not important. This options allows users for instance to report fiscal years starting from 1 March. We leave the default 20070101 14. Do you want to generate categories? If not needed, we suggest to generate them based on the data and do not generate a category for each day. No

We can see a new time dimension in the Dimension Map of the model:

The time dimension levels in the PowerCube are depicted below:

Define measures in the PowerPlay Transformer Model

After having all the dimension defined, we can proceed and create measures. When all dimensions and measures are set up, we will generate PowerCubes and open it in PowerPlay. Measures are numerical figures and come from the fact table. To add measures to the model, follow the instructions below: 1.Drag & Drop PRICE, QUANTITY and REVENUE columns from Data sources to the Measures window. 2. Double-click on the REVENUE measure and rename it to Gross Revenue . Then go to a Format tab to change the default number format to the #,##0 format with 2 Decimal places 3. Rename PRICE to Unit Price and apply the same format as for Gross Revenue. We also do not want to summarize the Unit Price measure because it does not make any business sense. To change the grouping calculation, go to Rollup tab and choose Average and set Regular Weight to Gross Revenue . 4. Set the #,##0 with 0 decimal places format for Quantity. 5. From the Data Sources window go to Facts and select DT. In the General tab change the Data class to Date and in the Time tab select the following Date input format: DMY

Measure properties for Unit price :

Define an output PowerCube

1. Go to the PowerCubes window and press Insert key 2. Put a name for the PowerCube: www.Cognos-BI.info Sales Analysis 3. Put the PowerCube file name: cognos-bi_analysis.pyi

 4. Create the PowerCube by clicking Run -> create PowerCubes and the model will become ready for exploration and analysis The final model structure is shown below:

The Cognos OLAP cube created from the model illustrated above:

Measures allocation
Measures Allocation is a techique that allows comparing data in a report even though it is grouped on a different level of detail. Usually it is required to perform allocation when part of the data is available only on a higher detail level than other data. In real-life, the measures that usually need to be allocated are fixed costs, forecasts, budget plan, discounts, rebates, customer bonus. This is because in most cases those figures are calculated on a monthly (for example COSTS), quarterly (CUSTOMER BONUS) or yearly (BUDGET) basis and very often the SALES reports are generated on a daily basis. It is also very important to remember that in some business cases allocation is logically inappropriate and may not make any sense. There are two ways of implementing allocation:

 Fixed allocation - which means that there is a constant value allocated to all levels in a dimension. Users may filter the report, drill up, down, slice and dice and the figure will remain constant. It may be sometimes confusing to the end users and before putting such a figure it is necessary to make clear to the end users that the number is constantly allocated and may not act as expected when analyzing data (especially when calculating summaries). Dynamic allocation (weighted, proportional) - values are allocated using subtotals of an another measure. First a percent of total of a allocation base measure is calculated and based on that figure, the corresponding values for allocated measure are calculated. The weighted type of allocation is often used in real-life datawarehouse environments.

Measures allocation real-life example in Cognos

In the following example we will analyze revenue, constant costs and variable costs. Our goal is to be able to compare revenue to fixed and variable costs in all the time dimension levels available. We will analyze the following measures from our tutorial business scenario: Gross revenue - the source data stores sales records on a daily basis. It will be handled automatically be cognos. Fixed costs - calculated per the whole year. The measure will be allocated. Variable costs - monthly detail level in a source data. This measure will be also allocated.

We will perform a constant value allocation to the FIXED COSTS in a YEAR dimension and dynamic allocation to the VARIABLE COSTS in a MONTH dimension (the allocation key is QUANTITY). To allocate a measure in PowerPlay Transformer go to a dimension level properties to the allocation tab and choose the corresponding option.

Fixed measure allocation:

Measure allocation by QUANTITY:

We see below that fixed costs have been allocated in a constant fashion:

Variable costs allocated dynamically by quantity:

Cognos Series 7 to IBM Cognos 8 BI products comparison

The table below shows the comparison of Cognos 7 and Cognos 8 BI tools and applications. The conclusion is simple: Cognos 8 is a completely different product than Cognos 7. Products comparison: Cognos Series 7 to IBM Cognos 8 BI Cognos series 7 IBM Cognos 8 BI Cognos PowerPlay Cognos PowerPlay Web BI Mobile Analysis (PowerPlay) Analysis Studio IBM Cognos 8 BI Analysis for Microsoft Excel ReportNet (Report Studio + Query Studio) Cognos Connection GO! Dashboard / Report Studio Data Manager Framework Manager Metrics Studio Event Studio Planning

PowerPlay Excel Impromptu Cognos WebPortal Visualizer DecisionStream PowerPlay Transformer / Impromptu Administrator Metrics Manager Notice cast Planning

Comments Desktop Query and Analysis tool Web-based multidimensional analysis MS Office / Spredsheet integration Reporting Web Portal Visual dashboards ETL & Data integration tool Modeling framework Scorecarding Business activity monitoring Planning, budgeting, forecasting

Cognos 7 Business Intelligence applications overview

Cognos Series 7
The main Cognos Series 7 Business Intelligence applications: Cognos PowerPlay - multidimensional analysis and OLAP reporting software. PowerPlay is available in a couple of versions: PowerPlay Web, PowerPlay Client and PowerPlay for Microsoft Excel. Cognos PowerPlay windows Client application has more features available than PowerPlay Web Explorer. Powerplay Enterpise Server is used to deploy cubes to the Cognos 7 WebPortal and also Cognos 8 Cognos PowerPlay Transformer is an application used to generate and update PowerCubes which are distributed to the end users PowerPlay Connect - an application which manages connections to Cognos OLAP source or any of the supported OLAP servers (Hyperion Essbase, IBM DB2 OLAP, Microsoft Analysis Services, local Microsoft .cub files, SAP BW) Cognos Impromptu - software for advanced reporting Metrics Manager - scorecarding software Notice Cast - event management application which runs in the background

The data model and flow of the data in Cognos is as follows: Analyze and define the data structure in source systems Use PowerPlay Transformer to connect to one or more sources of data and build multidimensional models Create the Cognos PowerCubes in Transformer and distribute those multidimensional models (and/or use PowerPlay Enterprise Server for distribution and to apply security to the OLAP cubes) View the distributed PowerCubes using Cognos PowerPlay for Windows, PowerPlay for Excel, Powerplay Web or Cognos Visualizer