Developing Business Intelligence with Data Science

Yuri Balasanov thinks of machine learning (ML) and artificial intelligence (AI) like hiking through mountains. When you climb a mountain, there is one summit in front of you, and that's what you're focused on. Once you reach the summit, though, you see a new summit you didn't know was there.  

That summit presents a new challenge. It also presents a new opportunity. 

The same, Balasanov said, is true for ML and AI. 

"We will continue being surprised, excited, and challenged again and again," said Balasanov, professor of instruction and deputy director of Northwestern Engineering's Master of Science in Machine Learning and Data Science (MLDS) program.  

Balasanov also teaches Data Science for Business Intelligence in Northwestern Engineering's Master of Science in Information Technology (MSIT) program. At least one-third of this year's class will focus on generative AI and large language models (LLMs).  

"This technology has a deep, transformative effect on every business," Balasanov said. "Data science and AI will continue changing every aspect of every professional activity so profoundly that getting trained for the change is critical for success in IT, as well as in any other domain." 

Balasanov provides that training in his Data Science for Business Intelligence course. This will be his third year teaching the class. His goal is to prepare future IT managers to work with data science teams so the two can collaboratively navigate their business' digital transformation. 

His secondary focus is helping MSIT students see how LLMs and generative AI can improve their productivity. To do that, Balasanov focuses on coding. 

"To understand what ML is about, one has to try coding it," he said. "Reading slides will not help. My MSIT students learn the same way as my data science students — by writing code in Python and fitting and fine-tuning the deep learning models using TensorFlow. The only difference is in the class time necessary for building up technical knowledge because MSIT students don't have the same depth of experience with ML and data science." 

Balasanov understands the challenges students may face with the steep learning curve. It's a challenge he faces himself in trying to keep the course content current. This year's class material varies greatly from the previous two years, primarily because the technology and use cases have changed so much in that time.  

To stay current, Balasanov has to keep learning himself. It's a tendency he hopes his students will emulate.  

"The most important takeaway, I hope, will be that the learning process must continue after the class is over," he said. "Some of the most important concepts that we will cover in this class did not exist a year or two ago. Next year I will add some new concepts to my curriculum that are unknown today. I hope that my students will leave the class with enough background to continue learning these new concepts on their own."