A significant contribution to the advancement of analytics

Bob Nydick has been teaching at the Villanova University School of Business for going on 30 years, and he’s a Wildcat to the core.

Nydick is passionate about the university in general and his teaching in particular, developing bonds with his students that have lasted through the years. "Maintaining contact is important to me," he says. "These are ongoing relationships."

In gratitude for the contributions his former students are now making in the world, Nydick, his wife, Susan, and their sons Rob and Greg decided they wanted to give back to Villanova in a substantial way. In the summer of 2014, the Nydicks made a $2 million commitment to the business school that will establish the Nydick Family Business Analytics Fund for undergraduate analytics education.

"I believe very strongly in the power of analytics and what it offers out in the world," Nydick says. "All the pieces for this gift fit together. We wanted to show our commitment to Villanova and its students, and we wanted to advance analytics."

Ensuring that analytics strikes a chord with students

Nydick is a devoted JMP® user. He believes that JMP – and particularly JMP Pro – offers a direct line to making an impact in business. "My PhD is in operations research, and in my opinion analytics is today doing what operations research should have been doing a long time ago – focusing on applications and ‘actionable insights," he says. "I think that term captures what the field of analytics is all about. What good is building these wonderful models if the results aren’t going to have real impact?"

With the right tools, students discover new applications as they go

Nydick’s undergraduate data mining course is designed around the book Building Better Models With JMP Pro, by Jim Grayson, Sam Gardner and JMP’s Mia Stephens. "Mia likes to say, ‘Get to know your data,’ and that’s where the course starts," Nydick explains. If, for example, there’s a significant amount of data missing, what do you do? "JMP Pro has the tools that allow you to recognize how extensive the missing data is, how much of an issue it is and what needs to be done to address that. 

"The same is true with outliers or correlated variables. The visualization capabilities of JMP allow you to easily identify them, and the students can then discuss how to proceed. JMP makes it very easy to identify and resolve those issues." The next step is to determine the most appropriate data mining model. Nydick teaches regression-based modeling, decision trees, bootstrap forests, boosted trees and neural networks.

At the end of the semester, he oversees a competition. All students receive the same real-world data set, usually about 50,000 records. Their task is to determine, for example, which subscribers are likely to switch cable or cellphone companies, or whom to target for a fundraising drive.

Students are evaluated on accuracy – the lowest misclassification rates – and scope – how many effective models they created to arrive at their predictions. "Some groups will come up with more than 300 models," Nydick says. They then choose which model to implement and present three actionable insights.

"I’m only teaching them a portion of the power of JMP. And what I find, especially with the graduate students, is that once they gain that base knowledge, they then explore and identify other features" – the Prediction Profiler, for example, which allows them to view how a prediction model changes as they alter the settings of individual factors. "If JMP wasn’t so intuitive, so easy to use, they wouldn’t be able to do things like that," Nydick says.

"I’m still learning JMP myself. I feel like I know a great deal about it, but I still discover nuances that end up in the course the next semester."

Big data, little models

Playing with new tools is fun. But Nydick underscores that it’s all about the advancement of those insights. "The key to me is the applications," he stresses. "You can do all sorts of exotic modeling, but if it doesn’t get presented properly, or the decision makers don’t have confidence in the results or don’t see the value in it, the work has had no effect."

Nydick has a real-world example he likes to share with his students. When he was launching his business, Leon Levine, founder of the Family Dollar franchise, would scout locations by scouring for oil stains in the parking lots. If the parking lot of a shopping center he was considering as a site for a new store was heavily stained with oil, it was a keeper.

Levine knew that many of Family Dollar’s customers had cars that leaked oil and that shopping centers that had evidence of oil stains were more likely to be successful locations. Over time, with careful analysis of a lot of data, Levine’s theory proved solid, and Nydick considers this a great example of what his Villanova colleague Matt Liberatore calls big data/little model.  

"Twenty years ago, we were creating a lot of big models with small amounts of data; the emphasis was more on the modeling and less on the data," Nydick says. "Today, modeling is no longer the emphasis of the process since we have so much data and that data provides key insights."

JMP tools like Fit Y by X, Graph Builder, Interactive Binning and the ability to transform variables are invaluable in helping identify key relationships, leading to improvements that help a business succeed, he says. "JMP is certainly playing a big role in the evolution of analytics, in why analytics is so successfully deployed today, and why I think it will be even more so five years from now," Nydick says. In sum: Actionable insights are now much more readily at hand.

Sending statistics-savvy graduates into the world

Thanks to the Nydick family’s gift, Villanova University will now be sending more graduates into the world with a firmer grasp of analytics and its universal applications. "The response has been beyond what we expected," Nydick says. The newly created analytics minor has become the university’s most popular, and the introduction of a major is on the immediate horizon.

For Bob Nydick, analytics is far more than an academic exercise. He’s helping lead his students to high-impact insights, imbuing them with enthusiasm for the potential of analytics. That’s a true gift.