Topic 10

A data warehouse is a database that is used for reporting and analysis. It is the MEGA BANK of information. It can take information from a wide variety of systems and business functions.

Example:

Operations
Sales
Marketing

Data from each system can be stored for easy recovery/mining.

Data warehousing is intended to cross platforms and be able to deliver a vast variety of data upon request.

Data mining pulls information that is pre-stages from the data warehouse for analysis.

_--------------------------------------------------------------------------------------------------------------------------------

These two systems work together quite nicely. The data mining tool that is being used is generally not able to communicate with a number of different systems. The tool is designed to talk to the data warehouse. It would be less efficient to have a tool that grabs if from all different systems on its own.

---------------------------------------------------------------------------------------------------------------------------------

A report can be created quite simply. First of all the report can be programed to run at a preset time so that it is available upon request. For example... at close of business the information can be gathered from the systems. This can take seconds to a few minutes or hours. Once the information has been pulled it can be run through an engine (program) that analyzes the data and presents it in a form that can be reviewed by a manager or executive. The whole idea is to push information to the correct individual..

----------------------------------------------------------------------------------------------------------------------------------

 The technical aspects are similar to the strategy. A program that is designed to open up bank accounts is not necessarily programed to send at the end of the day a report in a nicely formatted excel spreadsheet of a banker's performance. The technical side is about getting all forms of technology to talk to each other. It is important to take in mind the layers of data and filters that the data will be going through. Typically the less layers the better. It will generally yield quicker results and provide a better experience.