Presentation on theme: "End User Applications What to build. Metadata Management Project Management & Quality Assurance Source Systems Source Systems External Data Data Transformation."— Presentation transcript:
Metadata Management Project Management & Quality Assurance Source Systems Source Systems External Data Data Transformation Data Transformation Data Extraction Data Cleansing Data Integration Data Improvement Operational Data Store Operational Data Store Data Warehouse Data Warehouse OLAP Query Information Delivery Information Delivery Data Mining Operations & Systems Management Datamarts Enterprise Reporting Framework Architecture (Kerr McGee)
The Role of End User Applications Ad hoc power user Push-button knowledge worker Standard report customer Desktop tools / do-it-yourself queries Operational reporting environment End user applications
Application Template Provides the layout and structure of a report that is driven by parameters Users set parameters from pick lists or drop down menus
Cognos PowerPlay – web reports/slicing and dicing/data analysis, based on cubes.
Template Design Steps Identify an initial template set. 10 – 20 high value templates. Collect report candidates. What are the important report requests? Consolidate report categories. Based on data needs and type of report. Prioritize the list.
Common Report Types How’s business? What are the trends? What’s unusual? What is driving those exceptions? What if …? Make a business decision. Implement the decision.
Note Do not recreate the past. Old reports often do not reconcile Recreating old reports gains little
Template Management Navigation. How do you find the one you want? Shop standards. Naming standards for templates and columns Look and feel Template creation Definition and layout Match with dimensional model Review with users
Application Development Environments Web based Direct template access Structured/menu based template access Custom coded
Data Requirements It is hard to evaluate forms and templates without meaningful data. Problems to manage include: Meaningless descriptions Duplicate dimensions or dimension access keys Incorrect dimensional relationships Data not matching legacy reports. May be a legacy problem.
Managing the process Remember the milestones First demo25% System test with all modules25% Validated in simulated production25% Documented and delivered final system25%
Managing (cont.) Code is not self-documenting Use a bug reporting database or log Acknowledge bug reports from users and testers Place a high premium on testing and repairing