A concept, developed by W.G. Inmon, for the subject-oriented, integrated, time-dependent, and permanent collection of information in order to aid decision-making in a company. A data warehouse is a data collection that is isolated from the operational systems. It is used as a company-wide database for all areas of business to help with the decision-making process in a company.
In order to provide users with good quality, pertinent and timely information we often must access data from multiple source systems, reconcile differences between those systems, and provide a historic perspective.
In short, a data warehouse, a collection of all information for several sources that have been checked, cleaned and integrated to allow for users to have correct and accurate information.
To build a data warehouse, first decide what information the users require. Then create a database to hold the data, and populate it with the information from each source system. The source information system could be ERP, CRM, Access, Excel or any other source you fell contains important information. When the data collection is finished, we make the information available to business users through business intelligence (Cognos and BusinessObjects) and information management solutions.
Data Mart Design
A subject-specific subset of a data warehouse designed to meet the needs of a particular community of users and the security requirements of the organization. Data marts can be built for Finance, Sales, HR and any other functions to ensure that users in each area of the business see only the data that is relevant to them. Data marts also provide a performance benefit in that IT need only transport and provide access to a small block of data for each user community.
Data Marts are created to capture, "clean", transform, store and align data to support Business Intelligence (BI) requirements. They can be termed "staged" or additional databases in that they are distinct from the operational or line of business databases. Staged databases may hold 100MB of data or 100 Gig of data, but with the same objective - providing an architected and sustainable environment for data drawn from disparate data sources internal and external to the enterprise. A Data Mart typically covers a single subject or functional area whereas a Data Warehouse typically covers a complete enterprise.