Information Aggregation Application Classes

The Information Aggregation Usage page indicates the typical business needs for and uses of Information Aggregation applications, and the most likely type of BI application or tool to meet the need. The emphasis on this page is not on individual applications, but rather on a need for a particular class or set of classes of applications for each usage category. Generally, any of the business usage categories can use a mix of any or all of these Classes of Information Aggregation applications:

Executive Information Systems (EIS)

EIS systems provide high-level business information (balanced scorecard figures, key indicators, and so on) to upper management or executives. This information is often delivered to the user as a discrete item of work either automatically or on-demand. These systems focus on ease of use and integration with other elements of the executive's desktop, such as e-mail and calendaring.

High level reporting

This is typically performed by users (who work within the business community, for example business analysts) who run parameter-driven reporting queries or write ad hoc queries and reports against summarized data. For more sophisticated information, they would need more sophisticated data analysis or data discovery against detailed data and possibly rely on specialist support.

OLAP

Online Analytical Processing (OLAP) defines business information kept in a highly accessible form in which the data view can be easily changed. This is often termed "slice-and-dice," or "multi-dimensional" analysis. The information is usually summarized, aggregated, or derived, and frequently said to be stored in a "cube." Data is normally not the finest detail but offers high performance access to data for a large number of users. It is usually easy to drill down to a more detailed set of underpinning data.

Sophisticated data analysis

In this class, business users have preliminary ideas or leads about some element of the business. They then use appropriate tools (like SAS, for example) to perform a hypothesis-driven analysis of detailed data to discover relationships, confirm their impressions, and so on. The results of this work creates reports that run regularly in EIS and high level reporting. Intelligent Agent processes (as described in "Custom-built application" below) can also apply to this class of work.

Data discovery

This class differs from others in that the analysis is no longer driven directly by the user hypothesis. Instead, the user identifies a set of data to be analyzed and defines a problem domain. Then, a computing algorithm-driven analysis of the detailed data discovers associations, trends, patterns, and so on. This data mining discovers previously unknown affinities between data events, relationships between business events, and so on. Often data mining identifies key dimensions for OLAP cubes or new reporting parameters.

Custom-built application

Specialized business intelligence services (for example, using information from the data warehouse) for both internal and external users can be delivered through a custom-built application. A custom-built application performs a very specific set of functions, often including a mix of informational and operational aspects, integrated in a predetermined fashion as driven by the specialized needs of a class of business users.

Custom-built applications are increasingly commoditized. Vendors provide specialized BI functions in areas such as call center management, customer relationship management (CRM), supply chain management, and so on. These specialized systems provide a mix of operational and informational functions. They require careful planning to integrate with more generic BI functions.

Another class of custom-built application, an "Intelligent Agent" application, operates like a daemon process, looking for trigger events that initiate another process.

Knowledge Management

The topics of BI and Knowledge Management (KM) are increasingly merging. Though they have separate areas of competency, certain areas overlap and share common goals. Knowledge management is under consideration as a topic for a future application pattern.

Unstructured Data Solutions

The Data Integration::Population=Multi Step Gather, Population=Multi Step Process and Population=Multi Step Federated Gather application patterns can be used to create solutions that aggregate large stores of unstructured data gathered from the Internet. Such solutions retrieve and parse documents available on the Web and create an index of relevant documents that match a specified selection criteria input by a user.