It is easy to make generalizations in justifying a Business Intelligence System. Commonly used are:

• Saves money
• Helps enterprise to be more competitive
• Have informed decisions
• Improve productivity

and many others.

To face the critics of your BI system, quantify and provide specifics for your statements. For example, instead of just saying, it saves money, illustrate how the BI system saves money. If the pain point of your organization, for instance, is that one does not have a central repository of customer information so it takes accounting X amount of time creating an invoice because each time an invoice is created, one has to create a spreadsheet, getting information from multiple sources. It is effective to show how saving the time of accounting and billing the customer faster provides X amount of savings per month. As part of your justification, provide an estimate for the potential increase in receivables per month then multiply by the monetary amount of the average customer transaction.

A BI system that is aligned with business objectives and is able to maintain its economic justification gets buy-in and continuous support from the enterprise.

The concept of multidimensional cube is a good way to help users understand how they may want to query the multidimensional database or create OLAP reports. The dimensions of a cube are stored in a database table.

The data cube in the diagram below is composed of there dimensions: Customer, Product and Time.

This cube would allow query BY customer, BY time and BY product. Hence, sample query could be selecting a customer BY time and BY Product.

Additional dimensions (e.g. sales territory, sales person, etc) increase the size of the cube geometrically.

There has been a lot of concerns regarding the readiness of Open Source databases for the enterprise. Does this article: Sun Jumps On Open-Source Database Bandwagon To Boost Solaris prove that the enterprise has finally embraced Open Source databases?

Looking at the stats of Yackity Blog Blog, I noticed that First 3 Rules of Data Normalization for Newbies gets a lot of hits. So, I thought today I would write about a very basic concept - the Basic Constructs of Entity Relationship Diagrams (ERD).

The three basic constructs are:

1. Entity: A significant thing (place, person, concept, event) which information needs to be known or held
2. Attribute: A characteristic that serves to identify, quantify, classify or qualify an entity
3. Relationship: The way in which two entities are related

In ERD,
Entity is mapped to Table

Attribute is mapped to Column

Relationship is mapped to Foreign Key

Click to view original size

Getting input from the users increases the quality of a data model. Joint Application Development (JAD) sessions facilitate and accelerate the modeling process. During the JAD sessions, subject matter experts validate the model with respect to supported and identified functions and processes.

CRUD (Create Read Update Delete) Matrix is one technique to map the data model to the process model. The CRUD matrix helps identify:

• Missing Process : A business process exists that is not supported by the model; The data model needs to be adjusted to include the missing process
• Redundant Data Entity : A data entity exists that is not required to support a process

Think of CRUD in terms of how the data and process interact with each other. Using the simple example from the First 3 Rules of Data Normalization for Newbies, below is a simple example of a CRUD matrix:

Click to view original size

In this example, the process “Establish New Employee”, the entities (can be thought of as tables) of the model that are required to support the process are Employee, Employee Dependents, Warehouse (which warehouse will the new employee be assigned to).

Let us assume further that this model is only for employees and their training needs. Having an entity Product is a redundant entity. Of course, one can leave that entity in the model. However, it shows that for the scope of this particular model, it is not needed.