Posted on Monday, November 21st, 2005 at 9:15 am
Business intelligence depends on BI applications having access to properly prepared data. The normalized relational database formats used for transaction systems are excellent for rapidly reading, writing and querying data in a production environment. But as a native format, a transactional database isn’t well suited for BI.
For business intelligence apps, database programmers need to create data marts or data warehouses, which are a properly formatted amalgamation of all the key enterprise data. BI apps generate analytics or reports by querying and interpreting the contents of those data marts or data warehouses. (A data mart combines databases within a department; a data warehouse combines all strategic databases within the entire enterprise. For our purposes, a data mart is a small data warehouse.)
