Building A Data Warehouse (1,009 words)
Get Organized: Build a Data Warehouse
Now that you have this information, what do you do with it?
The objective is to centralize your customer data into a single repository, known as a data warehouse, where the appropriate decision support tools can be utilized for data analysis and trending. Initially, you will need to identify the data elements from the source databases that include their data definition. Then, you create a common set of metadata to relate the data to ensure consistency and integrity. Lastly, rules are developed using the metadata to perform extraction and transformation. These rules allow the data in the warehouse to be updated according to your business needs.
The structure of your data warehouse is primarily dependent on how you plan to use the data. If all the customer data in your warehouse should be available to all of your users, a Relational On-Line Analytical Processing (ROLAP) structure is recommended. If only subsets of the data need to be available to various departments in your company, then a Multi-Dimensional On-Line Analytical Processing (MOLAP) structure is preferred.
You will find that a data warehouse solution will facilitate sophisticated analyses across large volumes of data and enable you to visually drill up, drill down and drill through to your source data. You'll be able to play "what if" scenarios and automatically extract hidden predictive information from your databases. Most importantly, you will be able to visualize your data by finding a clever way to make sense of, and display, huge amounts of data stored in your corporate databases.
Eric Cohen is vice president of CACI Marketing Systems and Mark Bloom is vice president of CACI Enterprise Systems. CACI is a provider of demographic, consumer and business information systems and solutions. CACI has offices in Arlington, VA, (800) 292-2224, and La Jolla, CA, (800) 394-3690.