Business intelligence (BI) has long been a part of the corporate information systems landscape. Corporations learned long ago that with the proper BI infrastructure, they could increase profits, increase revenue and retain existing market share.

The profit motive is closely linked with the information that emanates from the BI infrastructure. With a properly architected infrastructure, the BI environment can handle difficult circumstances such as large amounts of data, ever-changing information requirements and a need to see data from many different perspectives. In addition, with a well-constructed BI environment, data can be presented in many different ways to the end user/analyst.

However, BI is not limited to the corporate world of information systems. BI is also important in government information processing circles. While profit is not a motive in government circles, fulfilling the information systems goals of the organization and living within the budget are. BI helps with both of those objectives.

Indeed, BI is no stranger to government. General Services Administration (GSA) has been using BI for years. According to Carmen Iannacone, team lead, database systems, GSA has used Sybase IQ for a variety of BI applications. GSA has used BI for determining trends in purchasing such as hot product segment areas, emerging products, agencies doing purchasing, seasonality in purchasing habits and the profile of the most active purchasers.

Sybase IQ has enabled GSA to find patterns that otherwise would not be apparent. For example, prior to the Iraq war, there was a rush to buy CamelBaks (individual hands-free hydration systems). Both branches of the service and individuals otherwise involved in the conflict made purchases. Through BI, GSA is able to detect and measure the trends.

Like their corporate counterparts, governments have large volumes of data and a need to see information in many ways. BI and its attendant architecture play naturally in those ways.

However, there is a special and important way that BI and its accompanying infrastructure play a role in government information systems. Government information systems are dominated by "stovepipes" – applications in which data is trapped. When the applications are built, the requirements are etched in stone. All data and processing are captured and managed internal to the application. There is no such thing as enterprise information in the face of stovepipe applications. Agency leadership is unable to look across the entire agency. The stovepipe applications are written so that they cannot be changed to any great extent. Because of their inflexible nature, data is local to the application and only to the application.

The center point of the infrastructure that supports BI is the data warehouse. It is through the data warehouse that stovepipe systems can be effectively addressed. In order to see how a data warehouse operates with stovepipe systems, consider the following dynamics. Upon request, the stovepipe systems move their data to the data warehouse. As the data is passed to the data warehouse, the data is integrated. A data model is traditionally used to provide an intellectual road map for the details of the integration of the data. As the data passes out of the stovepipes, it is transformed from the stovepipe form and structure into an agency-wide (i.e., enterprise-wide) form and structure. Multiple stovepipes contribute to this amalgamation of agency data. Once the data is transformed and merged from multiple sources, it loses its cast as stovepipe data and takes on the form of enterprise data.

The data that leaves the stovepipe environment and enters the data warehouse is granular or detailed. This atomic data is the most basic data the agency has; it is not effective for summary data to be taken from the stovepipe environment to be placed in the data warehouse.

Once the data warehouse is filled with granular data from the agency's stovepipe environment, there are many ways that the data can be used in traditional BI processing. In many regards, the integration of granular data from the stovepipe environment represents the liberation of the government agency from the tyranny of the stovepipe environment. If there is one way that BI contributes to the government's need for information systems, it is through providing an effective and immediate means of dealing with stovepipe systems.

Once the integrated granular data is organized and placed in the data warehouse, the data warehouse provides a basis for many different forms of BI processing. The data found in the government agency data warehouse is granular, historical and integrated. As such, the data is useful for many forms of governmental information processing.

One of the basic technologies supported by the data warehouse is online analytical processing (OLAP) or multidimensional technology. These traditional BI tools are attractive to the end user. The granular data is removed from the agency's enterprise data warehouse and reshaped into a multidimensional format. Or, the granular data found in the agency's enterprise warehouse can be used for decision support system (DSS) applications. Alternatively, the agency's enterprise granular data can be used for exploration processing and data mining. In a word, once the foundation is built properly, the world of BI is open to the agency.

There are many ways the granular enterprise-wide data found in the data warehouse can be used in governmental information processing. However, perhaps the most important usage of the granular data found in the BI agency enterprise data warehouse is as a basis for reconciliation when there are differences of opinion. Consider this: Two analysts are challenged by their leadership to produce similar analyses based on data found in the agency data warehouse. The two government analysts produce very different analyses with very different conclusions. Leadership is perplexed and wants to know why the analyses are so different.

If the analysts only have stovepipe systems supporting their conclusions, then there is no chance at any form of meaningful reconciliation. If a reconciliation can be done at all, it will take a long time. However, if there is a common warehouse supporting both analyses with data integrated down to the atomic level, even though there is a difference of opinion among analysts, there is still the strong chance of a timely and accurate reconciliation. In other words, the existence of a common enterprise-wide source of granular integrated, historical data provides a foundation from which to conduct reconciliation. For all of the other benefits of a data warehouse in a government environment, the potential for supporting reconciliation in the government information-processing environment is perhaps the most important reason for having the BI-oriented data warehouse.

However, there are other components of the BI environment which are of great benefit in the government information processing environment. Multidimensional processing is a legitimate component of the BI environment. With multidimensional processing, each department within the agency can look at the data that it needs in the format and structure peculiar to the department.

Normally, atomic data is fed to the multidimensional environment from the data warehouse. As the data passes to the multidimensional environment, the granular data is restructured into a format necessary for the multidimensional technology. This organization of data and processes is the same for the corporate and the government environment. Once in the multidimensional environment, the data can be manipulated in any number of ways. There is the possibility of drill-down processing and drill-across processing. The government leadership can look at the same data in many different ways and control how the data is presented.

It is noteworthy that if the multidimensional technology is connected directly to the stovepipe environment that permeates the government, it is not possible to conduct analysis across the agency using multidimensional processing. Multidimensional technology can only be used effectively where there is the foundation of data warehouse data and the liberation of the data from the stovepipe environment into multidimensional technology.

Another powerful form of BI processing is that of exploration processing and data mining. In exploration processing and data mining, detailed, historical data is examined statistically in order to find previously unknown patterns. In government information systems, these patterns may be for security, looking for telltale signs of a threat. In IRS taxation, the processing may look for signs of fraud. Census bureau processing may look for changes and shifts in population and demographics. Healthcare processing may look for abuse of the system, and so forth.

Exploration processing and data mining depend on detailed, historical data. The greater the amount of and the cleaner the data, the better. The stovepipe environment found in the classical government information-processing environment is hardly conducive to the effective execution of exploration and data mining activity. Data taken from the stovepipe environment must be integrated, cleansed and merged before it is fit for exploration and data mining. However, data taken from the data warehouse environment has already gone through the rigor of integration and transformation.

Because the data warehouse forms a solid foundation for exploration and data mining, the goals of BI are advanced in government information systems.

Another important aspect of BI processing is real-time processing. The ability to look at data in real time has many advantages. Additionally, in the BI environment, there is a structure known as the operational data store (ODS). The ODS is the place where integrated data resides and where online response is a possibility. If there is a need for very quick processing, the ODS responds.

When the ODS is fed from the stovepipe environment, the data must be cleansed and integrated. Furthermore, because of the limited resources found in the stovepipe environment, the feeding of the ODS becomes questionable. The stovepipe systems barely have enough power to run their normal processes. When a constant demand is made for draining data from the stovepipe environment, those resources simply cannot be spared.

Yet with a data warehouse, there is always the support of data feeding into the ODS. Typically, the data is amalgamated into a profile. The profile is then sent to the ODS and is available on a 24x7, 2/3-second access basis.

BI represents the first major step away from the stovepipe environment that has for so long plagued the government information systems environment.

Register or login for access to this item and much more

All Information Management content is archived after seven days.

Community members receive:
  • All recent and archived articles
  • Conference offers and updates
  • A full menu of enewsletter options
  • Web seminars, white papers, ebooks

Don't have an account? Register for Free Unlimited Access