Ch. 9: Business Intelligence Systems

Q1: Why do organizations need business intelligence?

Business Intelligence is information containing patterns, relationships, and trends that needs to be found and produced. Businesses use business intelligence systems to process this data to produce patterns, relationships, and other forms of information; and to deliver that information on a timely basis to users who need it.

Q2: What business intelligence systems are available?

A business intelligence system is an information system that employs business intelligence tools to produce and deliver information. The characteristics of a particular BI system depend on the tool in use.

A business intelligence tool is one of more computer programs that implement a particular BI technique. These tools can be categorized in three ways: reporting tools, data-mining tools, and knowledge-management tools.

Reporting tools are programs that read data from a variety of sources, process that data, format it into structured reports, and deliver those reports to the users who need them. Data are sorted and grouped, and simple totals and averages are calculated. These tools are primarily used for assessment.

Data-mining tools process data using statistical process data using statistical techniques, many of which are sophisticated and mathematically complex. Data mining involves searching for patterns and relationships among data. In most cases, data-mining tools are used to make predictions.

Knowledge-management tools are used to store employee knowledge and to make that knowledge available to employees, customers, vendors, auditors, and others who need it. It’s different from data-mining tools because the source of their data is human knowledge, rather than recorded facts and figures.

Q3: What are typical reporting applications?

 A reporting application is a BI application that inputs data from one or more sources and applies a reporting tool to that data to produce information. The resulting information is then delivered to users by a reporting system, which is a BI system that delivers reports to authorized users at appropriate times. Reporting tools produce information from data using five basic operations: sorting, grouping, calculating, filtering, and formatting.

RFM Analysis is a technique readily implemented using reporting tools, is used to analyze and rank customers according to their purchasing patterns.  RFM considers how recently (R) a customer has ordered, how frequently (F) a customer ordered, and how much money (M) the customer has spent. To produce an RFM score, the RFM reporting tool first sorts customer purchases records by the data of their most recent (R) purchase. The tool then re-sorts the customers based on how frequently (F) they ordered. Finally, the tool sorts the customers again according to the amount spent on their orders.

Online analytical processing (OLAP) is more generic than RFM. It provides the ability to sum, count, average, and perform other simple arithmetic operations on groups of data. The viewer of the report can change the format, hence the term online.  OLAP has measures and dimensions. A measure is the data item of interest. It is the item to be summed or averaged or otherwise processed in the OLAP report. Total sales, average sales, and average cost are examples of measures. A dimension is a characteristic of a measure. Purchases data, customer type, customer location, and sales region are all examples of dimensions.  With an OLAP report, it’s possible to drill down. An OLAP report is also known as an OLAP cube because some software products show these displays using three axes, like a cube in geometry. OLAP servers have been developed to perform OLAP analysis.

Q4: What are typical data-mining applications?

Data mining is the application of statistical techniques to find patterns and relationships among data for classification and prediction. It’s also known as knowledge discovery in databases (KDD).

With unsupervised data mining, analysts do not create a model or hypothesis before running the analysis. Instead, they apply the data-mining technique to the data and observe the results. With this method, analysts create hypotheses after the analysis, in order to explain the patterns found. One common technique of unsupervised data mining is cluster analysis where statistical techniques identify groups of entities that have similar characteristics. A common use is to find groups of similar customers from customer order and demographic data.

With supervised data mining, data miners develop a model prior to the analysis and apply statistical techniques to data to estimate parameters of the model.  One type of analysis is called regression analysis, which measures the impact of a set of variables on another variable. Neural networks are another popular technique used to predict values and make classifications such as “good prospect” or “poor prospect” customers.

A market-basket analysis is a data-mining technique for determining sales patterns. It shows the products that customers tend to buy together. In market-basket terminology, support is the probability that two items will be purchased together.  The ratio of confidence to the base probability of buying an item is called lift.

A decision tree is a hierarchical arrangement of criteria that predict a classification or a value. Decision-tree analyses are an unsupervised data-mining technique: the analyst sets up the computer program and provides the data to analyze, and the decision-tree program produces the tree. The basic idea of a tree is to select attributes that are most useful for classifying entities on some criterion. Then the data is input into the program. The program analyzes all of the attributes and selects an attribute that creates the most disparate groups. The program then examines other criteria to further subdivide.   

Q5: What is the purpose of data warehouses and data marts?

Many organizations choose to extract operational data into facilities called data warehouses and data marts, both of which prepare, store, and manage data specifically for data mining and other analyses.  This is because of problems with operational data.  The data could be dirty data, missing values, inconsistent, or the data is not integrated. Also, it could have the wrong granularity: too fine or not fine enough.  There could also be too much data: too many attributes, or too many data points.  A data warehouse is like a distributor in a supply chain. A data mart is a data collection, smaller than the warehouse that addresses a particular component or functional area of the business.

Q6: What are typical knowledge-management applications?

Knowledge management is the process of creating value from intellectual capital and sharing that knowledge with employees, managers, suppliers, customers, and others who need it. Whereas reporting and data mining are used to create new information from data, knowledge-management systems concern the sharing of knowledge that is known to exist.

Knowledge management is concerned with maximizing content use. Indexing is the most important function in knowledge management applications. Users needs an easily accessible and robust means of determining whether content they need exists, and if so, a link to obtain that content.

Real simple syndication (RSS) is a standard for subscribing to content sources. With RSS reader, you can subscribe to magazines, blogs, sites, and other content sources. In order to subscribe, the data source must provide what is termed an RSS feed, which means that the site posts changes according to one of the RSS standards.

Expert systems attempt to capture human expertise and put it into a format that can be used by non-experts. However, many of these systems were created in the late 80s and early 90s, and a few of them have been very successful.  There are three disadvantages: they are difficult and expensive to develop, they are difficult to maintain, and they have been unable to live up to the high expectations set by their name.

Q7: How are business intelligence applications delivered?

A business intelligence (BI) application server delivers those results in a variety of formats to devices for consumption by BI users. Some BI servers are simply web sites from which users can download, or pull, BI application results.  Another option is for the BI server to operate as a portal server, or as part of one.  Portal servers are like web servers except that they have a customizable user interface.

Q8: 2020?

Simple business intelligence systems like RFM and OLAP can successfully add value and even complicated and expensive data-mining applications can generate tremendous return if they are applied to appropriate problems and are well-designed and implemented.

  Kroenke, David. "Chapter 9: Business Intelligence Systems." Using MIS. Upper Saddle River, NJ: Prentice Hall, 2011. 318-347. Print.