Skip to main content

McCormick Making a Difference in New Era of Data Science

Key areas include analytics, optimization, new data science tools and materials design

soccer scene

Highlights

  • Data science mindset spreading across McCormick and Northwestern
  • Connection and collaboration critical to making data science thrive
  • Knowing how to extract information buried in data helps understand the world
  • Big data helps researchers create new knowledge
  • Employers beating down doors to hire graduates of data science program

EVANSTON, Ill. --- The competitive edge in the age of big data used to depend in large part on who had the data.

But today, with data growing exponentially and computational speed accelerating, the advantage goes to whoever can dig ever deeper, smarter and faster to make sense of the data.

“Knowing how to extract the useful information buried in the data can help us better understand the world and make informed decisions,” said Julio M. Ottino, dean of the McCormick School of Engineering and Applied Science.

For example, Ottino asked: What if you could analyze all the questions asked of Siri, the virtual personal assistant on Apple products that has become a staple of modern life? By inventing a system to extract patterns from the vast digital data and dig for certain information, one could gain insight about geographical regions, popular kinds of knowledge or any subset of frequently asked questions.

Along those lines, McCormick researchers are working with lots of big data sets and pushing data science forward in many ways. Ottino points to a number of key areas in data science research and education in which McCormick is particularly strong: data analytics, optimization, complexity, invention of new data science tools and the design of new materials.

Connection and collaboration are critical to making data science thrive at McCormick and the University, Ottino said. Researchers increasingly are forming interesting and fruitful partnerships across disciplines and schools, resulting in richer findings and insights.

“It’s about the mindset”

Central to efforts to advance data science at Northwestern, Luis Amaral is using his experience as co-director of the Northwestern Institute on Complex Systems (NICO) to increase communication across the University and build data science expertise and collaboration networks among faculty and students.

NICO has been very successful at connecting researchers interested in complexity science (the study of systems that display organization without a central organizing principle), which often involves big data. Amaral and his colleagues — including NICO co-director Brian Uzzi of the Kellogg School of Management — are using NICO’s infrastructure to help jumpstart the University’s data science initiatives.

Early efforts include supporting postdoctoral fellows and students already at Northwestern with data science experience as well as recruiting top postdocs and students with data science skills to the University. These postdocs and students will be instrumental in stimulating new collaborations among faculty.

“Data science is a way of thinking as well as a set of tools,” said Amaral, professor of chemical and biological engineering. “We want to help people at the University change their frame of mind for how they think about their data. These initiatives will help researchers create new knowledge, which is what scholarship is about.”

In his own research, Amaral has used publicly available big data sets to rank the success of professional soccer players, investigate the conditions that foster team creativity (conducted with Uzzi) and identify the most significant movies across the decades, to name just a few of his far-ranging data science studies.

“The movie data set we used was available at IMDB for anyone, but we were the first to use it in this way,” Amaral said. “Sometimes people are sitting on the data and don’t realize what they can do with it. Again, it’s about the mindset.”

Quest for Knowledge

Northwestern's high-performance computing network, Quest, is changing science and engineering at Northwestern.

Read more about Quest

Data science is a way of thinking as well as a set of tools.”

Luis Amaral
Co-director of the Northwestern Institute on Complex Systems (NICO)

Collaboration sparks new insights

Diego Klabjan and Moran Cerf both model the mindset. Using data science, they compared the brains of elite athletes in different sports performing similar physical biking exercises. Cerf, a professor of neuroscience and business at Kellogg, collected data about how brains function, measured as signals from the brains of subjects wired before and after exercise. Interested in exploring new ways to analyze the data, Cerf approached McCormick’s Klabjan, an expert in data analysis, for help.

Under Klabjan’s guidance, Liu Liu, a Ph.D. candidate in computer science, has been using data mining and statistical tests on the data to find similarities and differences in the functioning of the brains. While expected that brain regions close to each other would function similarly for all elite athletes, the researchers found that regions in the back of the brain interact more tightly than front regions. Strikingly, the brain’s reaction to pre- and post-exercising is very similar for all athletes in certain sports, e.g., water sports, while completely different in others, e.g., football.

“Working with Professor Cerf has been a great experience, with our complementary skills leading to great insights and important findings,” said Klabjan, professor of industrial engineering and management sciences.

Listen: Luis Amaral discusses how the university's cross-disciplinary approach enables the current transformation of data science.

Audio Transcript

Ties with industry advance knowledge

As director of the Master of Science in Analytics Program, Klabjan also is working to educate multi-dimensional data science experts for business. Students in the competitive program learn not only how to produce data-related software to add value to a company, but also how to talk to management about the findings from data analyses.

With faculty from McCormick, Kellogg, the Weinberg College of Arts and Sciences and industry, the four-year-old program covers computer science, statistics, business and communications. When students graduate, their deep knowledge of data science sets them apart, Klabjan said. Highly reputable companies, such as IBM, Google, Facebook and PayPal, are beating down the doors to hire them.

At the heart of data science is computer science, an area of strong growth at McCormick. Newly invented tools use computer algorithms and other methods to yield insights from often impenetrable big data. (Algorithms are step-by-step instructions, typically repeated many times, that allow a computer to solve a problem or help perform a function.)

Quill, one such tool — now used by Forbes magazine, The Big Ten Network and Mastercard — blends artificial intelligence with journalism. Born at Northwestern, Quill now helps organizations and companies more easily understand their data by turning it into a customized written story, about, for example, last night’s baseball game or an investment portfolio review.

Computer scientists Kristian Hammond and Larry Birnbaum, in collaboration with the Medill School of Journalism, Media, Integrated Marketing Communications, created the technology that became Quill. Six years ago, they co-founded Narrative Science, a startup company based on the patented platform.

In another close tie with industry, McCormick has combined its strength in materials science with data science to drive the new national materials genome initiative. The goal is to use big data, along with computations and experiments, to decrease by 50 percent the time required to create a new material and deploy it in industry. This materials design approach has significant implications for medicine, transportation, electronics and the environment.

Data As Art

Northwestern engineering and School of Art Institute students collaborate to visualize data.

Learn about the joint course

McCormick’s Peter W. Voorhees co-directs the $25 million Center for Hierarchical Materials Design, part of President Obama’s Materials Genome Initiative. He and his colleagues are creating enormous databases on various properties of materials and using the data to compute what properties to expect in a material if certain atoms are combined. This is the same approach Apple used to produce the Apple watch (made of three brand-new alloys, designed in part by McCormick alumnus Charles Kuehmann) in only two years.

“Databases — many larger than a terabyte — are absolutely essential for the materials design we are doing, which can result in huge savings to industry,” Voorhees said. “Data allows us to narrow down the enormous number of potential material compositions and become smarter about the ‘recipes’ we should try in experiments. Without data, we are back to the Edisonian approach of trial and error.”

Motivated by the application

To address the growing need for algorithms and models to process today’s voluminous data, McCormick recently launched the Center for Optimization and Statistical Learning. (Optimization is the process of finding the best design for a complex system or the best model for making predictions given a set of data.) The center draws together experts from computer science, industrial engineering, electrical engineering, business and medicine.

Northwestern researchers bring a practical problem from their field — planning effective radiation therapy, designing and operating a smart grid, improving natural language processing — and the center experts help develop a model and devise a solution algorithm, implemented in software. Surprisingly, the same types of optimization models and algorithms can be applied to data in a wide variety of areas.

“The application motivates us,” said David Morton, director of the center and professor of industrial engineering and management sciences. “Each large-scale problem results in a collaboration between the researcher with their individual expertise and center experts who have the methodological tools. Working together, we can make important advances that help inform decision-making.”

With today’s computational power generating dizzying amounts of data and astronomical feats of data crunching, Northwestern faculty and students increasingly are embracing data science and making new connections to take their research deeper.

“Data science is creating new possibilities in all disciplines,” Ottino said. “We truly are on the cusp of unimaginable new insights and scholarship that will change how we see the world.”

Read More

See how data science is impacting all fields and disciplines.

Back to top