Students use IBM Watson tech to build and balance big data

May 2017

Elizabeth Jenike

This past semester, the students in ISA 414 made some predictions - and they used some pretty cool technology to do it.

Arthur Carvalho, assistant professor of information systems and analytics, had his class spend the semester collecting, organizing and making sense out of data for various companies. They used the IBM Watson technology - the famous artificial intelligence program that has taken on Jeopardy contestants and which is now being used in fields such as health care to analyze patient data - to accurately describe trends, predict future events and prescribe what these companies should do to garner positive outcomes.

Essentially, they learned their way around big data sets. After all, “big data” is the name of the game in nearly every industry nowadays, and someone has to know how to read it.

Brand awareness through Facebook: A coffee story

All of the groups undertook interesting projects throughout the semester. One of the groups, comprised of senior management information systems major Edward Khaddam, senior accounting and analytics co-major Adam Levitt, and senior accountancy major Seth Levitt (pictured at right), analyzed Facebook data taken from several thousand posts that mentioned the Starbucks brand. The posts were gathered from before and after the company announced its plan to hire 10,000 refugees, in order to gauge, in real time, whether the plan was hurting Starbucks’ reputation.

All of these posts were fed into Watson. The technology, Adam said, was able to categorize Facebook posts according to whether they were positive, negative or neutral; and it also provided a breakdown of emotions (joy, sadness or disgust) for each post.

“Working with Watson was very exciting,” Adam said. “We were impressed with its accuracy, especially because it could recognize when posts were sarcastic or extreme.”

What did they find? Well, contrary to the assumption that the plan to hire refugees would create more negative sentiment in Facebook users, they found the opposite to be true. Positive sentiment increased toward Starbucks after the announcement, while stock price actually went down - showing that the public felt one way, while investors felt another.

Showing Netflix the money

Another group looked at the public response to Netflix original shows, along with their stock price in a given date range, in order to learn what the company was doing well and what it was doing poorly. The students - senior business economics and analytics double major Andrea Curtis, senior marketing and business analytics student Ashley Medaris, senior supply chain & operations management and analytics major Nicole Peters, and senior mathematics and analytics major Hailey Stafford - started this project in hopes to prescribe what actions would keep Netflix in a positive light.

“I certainly think that this project gave us experience working with advanced technologies in the future,” Hailey said. “Our use of Watson Natural Language Understanding was very similar to how a company would utilize this [software-as-a-service program]. We saw in our results that though the level of analysis the technology allowed was remarkable, the interpretation and actionable insights based on the analysis still required expert knowledge.”

Based on their collected and analyzed data, the students were able to provide Netflix with the idea that the company could use the sentiment knowledge, pair it with seasonal stock market forecasts and release trailers or announce show renewals when they anticipate a drop in the stock price. The idea being that positive sentiment surrounding trailers or renewal announcements could offset the anticipated decrease in stock price.

Carvalho had an opportunity to speak with representatives from IBM about the projects his class was undertaking, as the company was on campus to tour the Center for Analytics and Data Science, of which IBM is a sponsor. The representatives were surprised and pleased that Carvalho’s students were using Watson technology in their classes. About the Facebook project, they even said they wouldn’t have thought of that application of Watson.

Yet another reason our students are becoming more well-rounded, experienced business people every day.

Edward Khaddam, Adam Levitt, Seth Levitt