Dan Zaratsian, senior associate technical consultant in the analytics group, is an expert web-crawler—and wakeboarder—who works one-on-one with customers using math to mine sentiment from textual data. An electrical engineer with a mathematical mind, Dan thinks the text analytics he does is awesome—and so do we. Read on for a fantastic interview! And be sure to check out the rest of the SAS Loves Math series and the SAS STEM webpage for more information!
What do you do at SAS?
I support the sales and enablement of our text analytics suite of products. I work with customers to assess their business challenges and to derive value from their unstructured data through proof-of-concepts. Basically, customers give me their data, or they ask for data pulled from the web—web-crawling, which is my favorite—and I have to analyze this data, uncovering trends and patterns using our text analytics software. Some questions I usually get are: how do people feel about my company’s reputation? Are they strong proponents, or are they saying negative things? What’s the sentiment at a feature or product level? I translate emails, online data, social media data, blogs and forums, surveys, call centers… wherever there’s text, I jump in and analyze it.
Why do you prefer “web-crawling?”
First, it’s a challenge, because the web is rich with information. The web is perfect for me. I go to the New York Times and crawl articles. I go to credit card forums and look for fraud, people trading credit card information. I collect tweets pertaining to a specific company, event, or product. It’s interesting to explore the data, and it’s always dynamic.
How do you use math in your job?
Text analytics doesn’t sound like math, but it’s a core part of our technology and my day-to-day work. I rely on all kinds of different clustering algorithms and matrix reduction techniques that allow us to translate textual data into something meaningful. For example, I might look at the top twenty clusters that emerge from a collection of tweets. These clusters help us understand what people are talking about and what they are most interested in. This “enriched” data can then be merged with traditional mathematical techniques and models to further enhance overall predictive power and accuracy.
What are some of the cool things you’ve done with text analytics?
As a side project, I used math and some of our tools to pull in what people were saying about the Bellagio Hotel in Las Vegas. I love Vegas and I love casinos, so I wanted to know what people thought about one of my favorite ones. I used text analytics to crawl Twitter and I found out what people were saying. Not surprisingly, the sentiment was pretty positive, and there was a lot of talk about the Bellagio fountains, which are pretty cool.
Do you do a lot of work with social media?
I’d say over half of the work we do is tied to social media. Companies always want to hear what people are saying about them on Facebook or Twitter. I use Social Network Analysis—one of my favorite tools by far—to determine how people are related to one another, how often they converse, what their common traits are… this insight can help to review influential and often unseen patterns within the data.
What is your educational background?
I got my bachelor’s degree in electrical engineering from the University of Akron in Ohio. I worked as an engineer for three years prior to going to NC State’s Institute for Advanced Analytics, where I got my master’s degree in analytics. It’s the nation’s first master’s program in analytics and I really enjoyed it. I would highly recommend the program to anyone who’s interested. It’s only ten months long, but it’s a very intense, focused program filled with plenty of math and SAS coding.
What inspired you to become an engineer?
There was some family influence, because my uncle is an engineer. But it also just seemed natural to me. I knew I really liked math, so I went into computer engineering and switched to electrical engineering later. I eventually got a feel for what I liked and what I was good at, and that evolved into analytics—and I love the position I’m in now. Text analytics is awesome!
What about math appeals to you?
Well, first of all, I’m very analytical and I’m pretty meticulous. I’ve always been about recording things, from my health to my finances. What I like about math is that it helps me understand patterns and build upon those patterns to make decisions. I was never big on proofs or theories—I’m all about applied math.
Do you have a favorite mathematician?
Can I name a favorite artist? Leonardo Da Vinci. The reason I named an artist is because a lot of what I do with text involves turning it into “art.” People like to see pictures. We do it all the time with numbers (charts, plots, etc.), but text is a real challenge because it doesn’t lend itself naturally to those types of visualizations. I really admire artists because they take abstract thoughts and turn them into something that’s visually interesting. I try to do the same thing with abstract text.
What are your other hobbies and interests?
I love watersports. I’ve been wakeboarding my entire life. And waterskiing, barefooting, sailing, kite boarding… anything on the water. I also love to travel, and I’ve spent several weeks in Australia and Europe.
What kind of advice would you give to students studying math?
There’s always going to be the type of math that you like and the type of math you don’t. Even for me, I love math, but there are some things I just stay away from, things that are too unpractical for me—like proofs, for instance. But just because you run into certain types of math you don’t like, that shouldn’t deter you from being an engineer, or an analyst, or a mathematician.
Feel free to contact Dan about his interview by emailing him at dan.zaratsian@sas.com or commenting on his posts on the Text Frontier blog.
If you know someone at SAS who really loves math, nominate them for an interview! Just email me at dylan.sweetwood@sas.com.
