[June, 1997 Issue of the WICCA Basket, the newsletter of the Washington, D. C. area chapter of the Independent Computer Consultants Association, used by permission]

A major problem facing us all is the growing superabundance of data, piling up like snowdrifts all around us. It is clear to us at Nautilus Systems that the radical advances in data mining and data warehousing technologies enable us to alter our fundamental approaches to automation. We are now applying tools to automate analysis and extract knowledge, which goes far beyond our current use of automation to replace manual effort. What is also clear is that businesses need to do this to maintain their competitive edge.

The key is the integration of multiple disciplines: traditional database management technologies, very large databases, artificial intelligence (machine learning), statistics, geographic information systems, networking, communications, and Internet/intranet technologies. The new paradigm is emerging as methodologies mature supporting the development and implementation of data warehouses and the knowledge discovery and query techniques for exploiting them.

Nautilus Systems applies over thirty years of application development of artificial intelligence (AI), object-oriented technologies, very large databases, and communication networking.

Data mining is the use of knowledge discovery, pattern recognition, data analysis, and expert systems technology to automate the search for information locked up in (typically) very large databases. Although Nautilus Systems initially applied this expertise to acoustic analysis, these same techniques are the cutting edge of decision support systems, which we have implemented for sales and inventory analysis, database marketing, fraud detection, and financial prediction.

Pharmaceutical Sales and the Seasons

We conducted a market distribution study for an international pharmaceutical company producing both prescription and over-the-counter drugs. Using data mining and On-Line Analysis Processing (OLAP) techniques, Nautilus Systems was able to analyze millions of records that existed in different formats and different databases, and detect patterns that were not immediately apparent.

Nautilus Systems identified trends in the sales of allergy and cold medication by geographic area, by month. Looking at these trends on a map, it was immediately clear that sales of medication followed the allergy and cold seasons. Although this trend may sound obvious, the company managed its organization by regions and its product life cycle was not synchronized with the seasonal changes. It was being outsold – was its competition better able to address market timing?

It was clear that maximizing sales of each type of medication involved adjusting the timing of medicine reaching retail sales locations. Nautilus Systems applied workflow modeling to demonstrate the benefits of fine-tuning the entire production cycle to adjust the sales timing. Changes affected marketing, warehousing and distribution, the scheduling of pharmaceutical compounding, and the storage and delivery of compounding materials and packaging.

The company saved over $2 million per year by adjusting the timing of allergy and cold medication reaching retail sales locations. Nautilus Systems also developed automated extraction programs addressing data standardization, clean-up, and discrepancy reconciliation. These programs fed the Competitive Intelligence decision support database, enabling the company to continue to monitor the identified trends.

Dr. Jen Que Louie, president of Nautilus Systems, personally led this project. He may be contacted at jqlouie@nautilus-systems.com, or by telephone at (703) 385-8582. The Nautilus Systems web site contains numerous case studies of projects we have completed. The URL is http://www.nautilus-systems.com. We also have an electronic mailing list for company and data mining related news.