Data Warehouse
An alternative possibility is that customers can buy a package of tools and have the ability to build new applications using development tools and a platform. Custom applications, developed by customers themselves, are more likely to take advantage of the domain knowledge than vendors. Customers would prefer component based application development with graphical tools so that they don’t have to invest inordinate amount of resources in programming. A typical product of this nature which bundles a package of software tools and a development environment is Oracle’s 10g which combines business analytic and reporting tools, Oracle Discoverer for querying analysis and dashboard features; Oracle Spreadsheet Add-In, which allows direct access to Oracle’s online analytical processing (OLAP) from within Microsoft Excel spreadsheets; Oracle Warehouse Builder with extract, transform and load (ETL) capabilities; and Oracle BI Beans, a set of custom developing tools.
The advantage of custom applications is that they readily adapt to the existing infrastructure of the company including the data warehouse. In addition, companies have already implemented their data categories, dimensions and the data model and their customized applications will adapt better to this ensemble while packaged applications would require considerably greater adjustment within the enterprise including changes in the data warehouse.
The latest research, based on feedback from users, indicates that ease of integration is the single most important criterion for selection of Business Intelligence software. More than 75% of the users consider this to be the most important factor in the selection of BI software. The high level of significance of integration is illustrated by the case of Alaska Airlines which preferred Siebel Analytics for the ease of integration of the software in its heterogeneous environment especially the data warehouse. The metadata that came with Siebel Business Analytics is especially useful in the integration process.
Data warehouses have been the lynchpin of strategic applications of business intelligence; they enabled companies to coalesce information from disparate sources. In addition, the data warehouse stores data only after it has been cleaned for inconsistencies and when it conforms to standard definitions. Data is prepared for use in data warehouses by ETL (extraction, transformation and loading) tools which extract data from operational data stores, transform the data so that data definitions are consistent and duplicated data is removed and the output is loaded into data warehouses in line with its metadata. The data warehouse updates the same information. The downside of the data warehouse is that it can update information only when analytical functions are interrupted while data can be updated usually on weekends or beyond work hours. As the volume of data grows, the time spent on loading increases to an extent where it would conflict with data analysis functions. Also, real time analytics is really not feasible with the latencies that are inescapable with the data warehouse.