UA Professor and Students Mine for Data Nuggets

Dr. Michael Hardin has come up with a way for students to develop the
skills to use complex data-mining software

Dr. Michael Hardin has come up with a way for students to develop the skills to use complex data-mining software.

By Lauren Bocchino and Joanna Hutt

Maybe you have returned to a favorite commercial site, like Amazon.com, and been greeted with "Hi, Your Name! We have a new release on CD that you’re going to like. And how about that new thriller by John Grisham? You can pre-order right now."

No matter how jaded we computer users may think we are, this ability to recognize a consumer as an individual is still surprising, ... and ego enhancing.

Now consider a student who may be thinking of not returning to UA for his or her sophomore year, perhaps because of trouble with course work or family upheavals. Imagine how that student might feel to receive a letter from the registrar, saying "Dear Student’s Name: We want all our students to succeed at The University of Alabama. We have a plan to help you return and to increase your chances of success."

Collecting nuggets about consumers or students is certainly not new. That’s data. Recognizing and recording patterns from the mountains of data that businesses and institutions have collected is the key to marketing in a personal way, and marketing to individuals turns nuggets into a gold mine. That’s data mining.

A UA professor has come up with a way for students to develop the skills to use complex data-mining software -- software that turns mountains of data into intelligent information that users can mine and put to best advantage. These students will in turn be able to develop ideas for how UA can best use the intelligent information gleaned from data it mines.

UA’s Culverhouse College of Commerce and Business Administration (C&BA) has teamed with the SAS Institute, a market leader in business intelligence and data warehousing, to create a four-course, credit-carrying sequence in data mining, which many observers think may be the next "hot" area of "business analysis." Dr. Michael Hardin, professor of statistics in the department of information systems, statistics, and management science, was influential in bringing the program to the University.

"I have been working in the area of data mining since 1995 and have observed the tremendous need for individuals who could assist business and industry in successfully building predictive models and other data mining tasks. Such skills must be developed by hands-on experience with large data sets and leading edge technology," Hardin said.

However, he said, to gain the experience students need access to the software provided by SAS. In turn SAS will have a trained group of students familiar with the software who want to continue using it after they graduate and begin their careers.

"By partnering with a leading vendor in this area, I believe we can more quickly provide an educational experience for our students to meet this marketplace demand. SAS, with the Enterprise Miner, is the industry leader in data mining software," Hardin said.

Dr. Roger Thompson, University registrar, bought into Hardin’s plan by providing the mountain of data. "Data mining will aid in recruitment and retention efforts. The software does problem solving based on research, and the models the students will create will help UA to solve problems," he said.

Fifteen C&BA students will eventually make presentations on how the Office of Academic Records and the registrar can best use the mining techniques they learn in the course work. Thompson even put a member of his staff, Teri Terry, coordinator of Student Information Systems (SIS) and data resources, in the courses to learn with the students.

Enterprise Miner allows students to compile the same amount of data in one afternoon that would take five to six months to compile without the software.

"Our sequence of courses will offer students the advantages of working in project teams using real data and our courses will be project oriented," Hardin said. "Because of the structure of the sequence, students from across the business school will be able to participate in the program, bringing an interdisciplinary perspective to the courses and projects."

While it is often several years before a course can be offered in an emerging area such as data mining, the SAS partnership provides the College a jump-start in the ability to offer these courses, Hardin said.

Hardin, who has also taught basic SAS programming overviews, said a class in data mining could cost $1,200 per student outside of the University. By offering it at the University, however, Hardin said, a center of excellence known for producing top quality data miners can be established. As an added incentive, SAS will award students who complete the four-class course a data mining certification, Dr. Hardin said.

The first set of courses was available this past semester and included DATA STEP, PROC SQL, the SAS macro facility, the Output Delivery System, and data warehousing. The final two courses that will complete the sequence in data mining include both theoretical basis and practical applications of common data mining and analytical tools such as neural networks, decision trees, cluster analysis, and association analysis.

"This partnership will enrich several of the College’s nationally ranked programs, foster collaboration among our business faculty and enable us to gather deeper insights into the problems modern businesses must solve." Hardin said. He added that in addition to the short-term goals the SAS partnership will provide, the University would also benefit long-term.

Thompson agrees. "This is a team effort," he said. "Who ends up winning? We all do -- the students, the College and The University of Alabama. And Dr. Hardin made it happen."