by Amy K. Lavender

On Friday, the University of West Georgia’s Richards College of Business welcomed experts in the field of data analytics for the fifth annual SAS Analytics Summit. Students as well as faculty, staff, community members and area professionals were invited to the event, which showcases the variety of applications for large data analytics across multiple fields. 

Student showcasing data research
Students showcase their data research during the fifth annual SAS Summit on Friday, April 21, at UWG.

Dr. Joey Smith, professor and chair of the UWG economics department led the event and introduced the speakers.

“First of all, I have to thank SAS and David Johnson, without whom this event wouldn’t be possible,” Smith said. “He and SAS have worked together to give us the best software available these days as well as the opportunity to talk about how large data is processed and all the interesting things we can do with it.”

The first speaker of the day was Andre´ de Waal, an analytical consultant with SAS, who also spoke at last year’s summit. De Wall spoke about the field of analytics and its various applications.

“I can honestly say that I am in the right area, in the right place and at the right time because advanced analytics is exploding and everyone wants to get into this field right now.” 

He explained advanced analytics, or “big data,” as taking large data sets (think billions of observations) and making tools that can analyze the data and provide users with a quick response.

“SAS visual analytics tries, basically, to democratize analytics,” de Waal explained. “Anybody in this room who needs to work with large sets of data should be able to benefit from SAS analytics, whether you’re trying to find out how customers are using your product or what kind of sentiment your company or product is generating.” 

The new frontier of analytics, de Waal said, is text analytics, which contain no numbers. As an example, he brought back his data from the first presidential debate between Donald Trump and Hillary Clinton, which was all text-based.

“If you can harness this program and use it for unstructured data, there’s no end to its applications; from medical records to customer service records, these all are text-based, there are no numbers,” de Waal explained, “but any of these fields can benefit from these analytics.”

The second keynote speaker of the event was Jorge Silva, a senior research statistician developer for SAS, who spoke on unstructured data through machine learning.

“Unstructured data comes in many forms, but we’re going to talk today about sparse data,” Silva said.

Using the example of Netflix, Silva demonstrated how the company had to create an algorithm to recommend movies to customers based on very little data.

“Netflix even offered a $1 million prize to anyone who could solve this problem, because it was a real problem. They had to find a way to make movie recommendations that were relevant,” Silva explained. “They have hundreds of millions of customers and tens of thousands of movies, and the only data they have is movie ratings. The problem is, no one rates the movies. In fact, less than 1 percent of their customers rate movies.”

The only other data the company had was transactional data—records of which movies each customer watched.

“This is extremely sparse data,” Silva explained. “So we have to come up with a clever way to take a matrix and insert the data … that gives you a method for predicting ratings that you don’t observe.”

Silva elaborated on ways to take a factorization model and extrapolate the data you are searching for from the sparse data on hand as well as determine if the model has value.

After a short break, during which students displayed their data research projects, the summit resumed with a panel discussion, featuring David Johnson, a partner with Cane Bay Partners VI, LLLP; Salome Deka, a senior analyst for the Women, Infants and Children (WIC) Program at the Georgia Department of Public Health; Laureano Gomez, a senior analyst with Epsilon; and Elizabeth J. Wang, a senior statistician at Epsilon.

The panel talked about their careers, how they earned their current positions and what specifically they do in their jobs. They shared tips with the audience about how to confront difficult tasks as well as how to know when you have a “good enough” answer.

“It’s tough in this field to not be a perfectionist,” explained Johnson. “You do as good as you can as quick as you can, but if it’s going to take an extra two months to make it perfect, you have to look at it from a practical standpoint and determine if it’s worth it long term—especially with serious applications, like health care.”

The most important message the panel shared with attendees was to continue learning new things and gain value from their work.

“Even after all these years, I still ask myself, ‘What do we know?, How can we relate that knowledge to the problem? What did we learn here?’ But most importantly,” Gomez said. “I encourage you to never stop learning.”

Posted on April 24, 2017