Southern California Gas Company
The Southern California Gas Company is using SAS software as a strategic marketing tool. The company maintains a data mart called the Customer Marketing Information Database that contains internal billing and order data along with external demographic data. According to the company, it has saved hundreds of thousands of dollars by identifying and discarding ineffective marketing practices.
WebWatcher
Despite the best effort of Web designers, we all have had the experience of not being able to find a certain Web page we want. A bad design for a commercial Web site obviously means the loss of customers. One challenge for the data-mining community has been the creation of “adaptive Web sites”; Web sites that automatically improve their organization and presentation by learning from user-access patterns. One early attempt is WebWatcher, an operational tour guide for the WWW. It learns to predict what links users will follow on a particular page, highlight the links along the way, and learn from experience to improve its advice-giving skills. The prediction is based on many previous access patterns and the current user’s stated interests. It has also been reported that Microsoft is to include in its electronic-commerce system a feature called Intelligent Cross Sell that can be used to analyze the activity of shoppers on a Web site and automatically adapt the site to that user’s preferences.
AbitibiBowater Inc. (Canada)
AbitibiBowater Inc. is a pulp and paper manufacturer headquartered in Montreal, Quebec, Canada. The pulp and paper, a key component of the forest products industry, is a major contributor to Canada’s economy. In addition to market pulp, the sector produces newsprint, specialty papers, paperboard, building board and other paper products. It is the largest industrial energy consumer, representing 23% of industrial energy consumption in Canada. AbitibiBowater Inc. used data-mining techniques to detect a period of high performance and reduce energy consumption in the paper making process, so that they recognized that lower temporary consumption is caused by the reduced set point for chip preheating and cleaning of the heating tower on the reject refiners. AbitibiBowater Inc. was able to reproduce the process conditions required to maintain steam recovery. This has saved AbitibiBowater 200 gigajoules1 daily—the equivalent of $600,000 a year. [Head Up CIPEC (Canadian Industry Program for Energy Conservation) new letter: Aug. 15, 2009 Vol. XIII, No.15]
eHarmony
The eHarmony dating service, which rather than matching prospective partners on the basis of their stated preferences, uses statistical analysis to match prospective partners, based on a 29-parameter model derived from 5000 successful marriages. Its competitors such as Perfectmatch use different models, such as the Jungian Meyers-Briggs personality typing technique to parameterize individuals entered into their database. It is worth observing that while the process of matching partners may amount to little more than data retrieval using some complex set of rules, the process of determining what these rules need to be involves often complex knowledge discovery and mining techniques.
The maintenance of military platforms
Another area where data-mining techniques offer promising gains in efficiency is in the maintenance of military platforms. Good and analytically based maintenance programs, with the Amberley Ageing Aircraft Program for the F-111 a good example, systematically analyze component failure statistics to identify components with wear out or other failure rate problems. They can then be removed from the fleet by replacement with new or reengineered and thus more reliable components. This type of analysis is a simple rule-based approach, where the rule is simply the frequency of faults in specific components.
B.6 PITFALLS OF DATA MINING
Despite the above and many other success stories often presented by vendors and consultants to show the benefits that data mining provides, this technology has several pitfalls. When used improperly, data mining can generate lots of “garbage.” As one professor from MIT pointed out: “Given enough time, enough attempts, and enough imagination, almost any set of data can be teased out of any conclusion.” David J. Lainweber, managing director of First Quadrant Corp. in Pasadena, California, gives an example of the pitfalls of data mining. Working with a United Nations data set, he found that historically, butter production in Bangladesh is the single best predictor of the Standard & Poor’s 500-stock index. This example is similar to another absurd correlation that is heard yearly around Super Bowl time—a win by the NFC team implies a rise in stock prices. Peter Coy, Business Week’s associate economics editor, warns of four pitfalls in data mining:
1. It is tempting to develop a theory to fit an oddity found in the data.
2. One can find evidence to support any preconception if you let the computer churn long enough.
3. A finding makes more sense if there is a plausible theory for it. But a beguiling story can disguise weaknesses in the data.
4. The more factors or features in a data set the computer considers, the more likely the program will find a relationship, valid or not.
It is crucial to realize that data mining can involve a great deal of planning and preparation. Just having a large amount of data alone is no guarantee of the success of a data-mining project. In the words of one senior product manager from Oracle: “Be prepared to generate a lot of garbage until you hit something that is actionable and meaningful for your business.”
This appendix is certainly not an inclusive list of all data-mining activities, but it does provide examples of how data-mining technology is employed
