Other Types of Business Intelligence - dummies

By Thomas C. Hammergren

Alas, the neat, organized model that has four different types of business intelligence categories (querying and reporting, business analysis [OLAP], data mining, and dashboards and scorecards) can be expanded for more complex applications. For example, an OLAP or dashboard tool might have geographical information system (GIS) capabilities — or it might not.

As shown in this figure, business intelligence includes horizontal categories that span some or all of the vertical types.


Statistical processing

Most business intelligence tools do rudimentary statistical processing, such as averages, summaries, maximum values, minimum values, and standard deviations.

Think back, though, to that statistics class you endured in college. (If statistics class was easy for you, more power to you.) Remember z-scores? Chi-square tests? Poisson distributions? Some folks in the real world really use those concepts (and many other related concepts that most of us left way behind) as part of their jobs.

These people always want statistical functionality integrated into a business intelligence tool so that they don’t have to save results from one tool into an intermediate storage facility (such as a Microsoft Excel spreadsheet) and import that data into a statistical tool for processing. Tool integration is a good thing!

Additionally, a large part of some data mining processing is based on heavy stats (advanced statistics, such as probabilities). Some environments include simulation and gaming capabilities that can identify and test various outcomes, and try to handle sophisticated what-if processing based on real data, rather than on assumptions and hypotheses.

Geographical information systems

Geographical information systems (GISs) make up a category of business intelligence functionality that spans multiple categories. The simplest way to understand GIS technology is to consider the concept of maps.

You could, for example, do querying in a tabular manner and show product sales by country. Those sales are divided into product sales by territory and then broken down into product sales by increasingly smaller groupings, down to product sales by department in each store, for example.

One business planner made the profound statement, “Our business is geographically based — why can’t I see our key information on a map?” GISs do just that: They provide a way of viewing information based on presenting that information by using maps.

Suppose that all countries colored in red indicate that sales revenue is lower in the most recent quarter than it was in the preceding quarter. If you click your mouse on any of those countries, another map appears, which is divided into sales territories and again color-coded. Clicking on a red-colored territory displays another map (a U.S. map by state, for example) with additional levels of detail.

When you use GIS technology, data isn’t just managed in a hierarchy such as this:


The data is also managed spatially — in a manner that’s sensitive to on-screen layout. For example, when you click a map of Pennsylvania, the GIS would “know” that you want to see sales in Pittsburgh, without your having to type or otherwise select that city from a drop-down list or other on-screen control.


The capability to mix and match information and presentation services, such as those found on the Web, is a simple example of a mash-up. Companies such as Google, Yahoo!, and Microsoft are driving much of the technology platforms for future applications and provide key technologies in the area of mash-ups.

Therefore, the example for GIS could actually take your standard database query and present that data by using Microsoft’s Virtual Earth technology. Furthermore, you might want to combine that information with external benchmark information to give you greater insight.

For example, if your marketing team felt that a new product would sell well to those individuals who have a median income of $50,000, you could plot census data with your sales performance data on a map provided by Google, Yahoo!, or Microsoft. How cool is that?

Although few companies are using mash-ups today, mash-ups will likely emerge as the future presentation vehicle. Current mash-ups combine external data, such as information that you can find on Craigslist or Amazon, with external visualization tools, such as Google documents or maps.

Business intelligence applications

While the business intelligence and data warehousing markets mature, large vendors of commercial off-the-shelf software, as well as specialty boutique firms, have started offering out-of-the-box business intelligence applications. These solutions typically are targeted at a functional area of a business, such as finance, supply chain, or customer relationships.

The solutions combine all aspects of subject area data content that has an associated database and set of business intelligence tools. These business intelligence applications can help solve 60 to 70 percent of issues you might commonly find in a functional area of your business.

The application comes with a set of predefined content, views, or reports that enable you to simply hook up the application to your current systems and go. Vendors such as SAP and Oracle dominate this area because they also control much of the “run the business” components required as data sources for the “monitor the business” applications that they provide.