Open Framework, Information Management Strategy & Collaborative Governance | Data & Social Methodology - MIKE2.0 Methodology
Wiki Home
Collapse Expand Close

Members
Collapse Expand Close

To join, please contact us.

Improve MIKE 2.0
Collapse Expand Close
Need somewhere to start? How about the most wanted pages; or the pages we know need more work; or even the stub that somebody else has started, but hasn't been able to finish. Or create a ticket for any issues you have found.

Business Intelligence Design

From MIKE2.0 Methodology

Share/Save/Bookmark
Jump to: navigation, search

Content Model Relationship

Contents

Activity: Business Intelligence Design

Objective. The Business Intelligence Design converts earlier process specifications of the Business Intelligence prototype into detailed design specifications required for the increment. The processes considered here follow-on from prototyping in the preceding phase. As stated then, such processes are not concerned with data acquisition, data transformation, or information presentation activities. This activity may consist of designing certain functions using an OLAP or data mining tool; 3 levels of users normally can be distinguished from within an organisation:

  • Business users: Chose to read the standard reports without user input parameters (static).
  • Investigative users: Chose to parameterise existing reports (slightly dynamic)
  • Power users: Create their own reports and queries or adapt existing reports (dynamic)

On average 70% of all users can be categorised as business users, so that a standard solution of distributing static reports can suffice for them.

Major Deliverables
  • Reporting Design (more traditional-type reports) for business users
  • Ad-Hoc Access Design
  • OLAP Design
  • Data Mining Design
Tasks

Task: Design Reports

Objective: Reporting Design applies to the more traditional design of static reports to support the Business Intelligence environment. It builds on the initial Report design document and prototype work that was done interactively with the users in Phase 3 to provide the ’hardened’ design for a production system.


Input:

  • Detailed Business Requirements for Increment
  • Initial Report Design
  • Target Logical Data Model and Physical Data Model


Output:

  • Detailed Report Design Deliverable Template
  • Detailed Report Design

Task: Design Model for Ad-Hoc Analysis

Objective: If users require ad hoc information access, it is common to provide a richer, abstracted environment from which they can access the information. Vendors have different approaches to this problem, but generally a well-defined metadata environment is what facilitates access to information and allows investigative and power users to directly answer the questions they have, complementing the information in existing reports.


Input:

  • Detailed Business Requirements for Increment
  • Initial Report Design
  • Target Logical Data Model and Physical Data Model
  • Integrated metadata into target environment


Output:

  • Model Designed for Ad-Hoc Analysis Deliverable Template
  • Model Designed for Ad-Hoc Analysis

Task: Design OLAP Model

Objective: OLAP (Online Analytical Processing) tools provide the capability to quickly analyse and report on multidimensional or time-oriented data, using techniques such as drill-down, drill-up, drill-across and visualisation. This would be a specific technology approach for implementing a solution for Ad-Hoc analysis. Multi-dimensional OLAP tools can offer advanced functionality in terms of executive information analysis and decision support, oftentimes providing a richer approach for ad hoc analysis of information. Many OLAP vendors have quite different ways of implementing solutions although in some cases the terminology differences mask fairly similar technology approaches.


Input:

  • Detailed Business Requirements for Increment
  • Initial Report Design
  • Target Logical Data Model and Physical Data Model
  • Integrated metadata into target environment


Output:

  • OLAP Model Designed for Ad-Hoc Analysis Deliverable Template
  • OLAP Model Designed for Ad-Hoc Analysis

Task: Design Data Mining Procedures

Objective: The process for Data Mining is different from that for decisions support: whereas a DSS system provides answers to questions, Data Mining is about discovering new questions and the associated answers. Data Mining is a discovery-driven process involving the use of detailed and historical data. Its applications are most commonly focused around Marketing Analystics, Fraud Detection, Risk Management and Finance. Common Data Mining operations as defined in the Business Intelligence Roadmap [1] include the following:

  • Predictive and Classification Modelling: This data mining operation is used to forecast a particular event. It assumes the analyst has the specific question that he or she wants to ask
  • Link Analysis: The data mining operation is a collection of mathematical algorithms and visualisation techniques that identify and visually present links between individual records in a database.
  • Database Segmentation: This data mining operation is a set of algorithms that group similar records into homogeneous segments. Segmentation is commonly applied technique for information platforms, not necessarily only being applied when doing data mining.
  • Deviation Detection: This data mining operation is a set of algorithms that look for records that fall outside some expectation or norm and then suggests reasons for the anomalies.

Data Mining, like Data Re-Engineering, often follows an iterative development process based on the \’80-20 rule\’. The standout findings are discovered in the early stages and the discovery process which produces the analytical data model is built out over time.


Input:

  • Detailed Business Requirements for Increment
  • Initial Report Design
  • Target Data Model (Logical and Physical)
  • Integrated metadata into target environment
  • data-primed testing environment


Output:

  • Analytical Data Model to be used within the Data Mining Tool Deliverable Template
  • Analytical Data Model to be used within the Data Mining Tool

Role: Information Architect

Role: BI Application Development Lead

Role: BI Application Developer

Yellow Flags

  • Frequent changes required to analytical data model
  • Representative data sets are not available for prototyping

References

  1. Business Intelligence Roadmap: The Complete Project Lifecycle for Decision-Support Applications. Larissa T. Moss, Shaku Atre (Addison-Wesley Professional).
Wiki Contributors
Collapse Expand Close

View more contributors