Putting It All Together
“As a kid I loved puzzles,” says Jason Okui. “Today I still love puzzles, but now the puzzle pieces are business data. The tools are different, but my passion for fitting the pieces together is the same.” Okui fits those pieces together for software giant Adobe, where he directs business operations and strategy, a promotion he earned after completing an Online Master’s in Predictive Analytics (MSPA) degree at Northwestern in 2016.
“The biggest benefit of the program for me was that it gave me a deeper understanding of how predictive models work and how they can be applied. Just as important, it allowed me to communicate what a predictive analysis means,” says Okui. “For all the sophisticated modeling you can do, if you can’t explain the model to the business world, it won’t be implemented.”
Okui relies on analogies to cut through what might otherwise sound like statistical mumbo jumbo. “Pizza comes in hundreds of varieties and combinations, but most pizzas share three ingredients: dough, tomato sauce, and cheese,” he offers as an analogy. “If I can explain a predictive model in terms of what best captures the essence of pizza, there’s a good chance you will be more comfortable with the fancy statistical things we are doing and trust those results more.”
Winnowing down variables was exactly what Okui and three of his MSPA classmates did for their capstone project in the program. The team’s challenge was to develop a model to predict where Airbnb users would book their first international trips.
“We started with 566 variables and narrowed them down to the 24 that mattered most,” says Okui. The team communicated their findings in a PowerPoint presentation, with Okui functioning as project manager: “We wanted to tell a story, to demonstrate how people become inspired about travel. This wasn’t a typical academic paper. We treated it as a real business project, and it became a template for how I communicate findings at work.”
The team entered their model in an open-data competition on Kaggle, a platform for predictive modeling, where it ranked in the top 3 percent on the site’s leaderboard.
Electives in Marketing Analytics and Advanced Modeling Techniques have served Okui well in his career. “I chose the program for its well-rounded curriculum,” he says. “One of my biggest regrets in college was not taking enough statistics. The program showed me the nuances. Predictive analytics is an art, not cold math. A lot of creativity goes into it. Even within a statistic, there’s leeway in the way I transform the data.”
Okui lives with his wife and two children in California’s Castro Valley area, only a mile from where he grew up. He started the MSPA program in 2012, just after he landed his job at Adobe. “The online component was critical. It allowed me to schedule my studies so that I could have dinner with my family every night. Now that I’ve graduated, I finally have time to take swing dance lessons with my wife.”