MSIS student Rashnil Chaturvedi and teammates compete in final round of international analytics contest

April 19, 2017

Traveling to Las Vegas earlier this month, Rashnil Chaturvedi had his sights set on a big prize, but he and his teammates were not relying on luck. They brought months of in-depth data analysis, a sophisticated modeling approach, and a polished presentation to compete in the INFORMS O.R. and Analytics Student Team Competition.

Rashnil Chaturvedi, Scott Smith, and Aniish Sridhar in Las Vegas for the final round
of the INFORMS O.R. and Analytics Student Team Competition.

Chaturvedi, a Master of Science in Information Science (MSIS) student at the UNC School of Information and Library Science (SILS), learned about the competition from connections he had made through his minor in statistics and operations research (STOR). Two master’s students in the STOR program, Aniish Sridhar and Scott Smith, received an invitation from INFORMS last fall and encouraged Chaturvedi to form a team with them. They and other teams from all over the world tackled a complex problem provided by Syngenta, a seed biotech company.

“Syngenta gave us their raw data on around 16,000 soy bean varieties that they had tested at 151 locations,” Chaturvedi explained. “They wanted us to use the data to create a modeling technique that would help them with earlier detection of varieties that should be commercialized. We spent two months just understanding the data because it was so complex with over six years of information.”

The team’s written analysis earned them one of just eight coveted spots in the final round of the competition. INFORMS flew them to Las Vegas to present their solution to a panel of judges at the INFORMS Business Analytics Conference held April 2-4. Though the UNC team did not place first, Chaturvedi said the experience (not to mention the free trip and $500 honorable mention prize) was well worth the effort.

“Before starting this project, I didn’t know you could apply this kind of modeling technique on such a granular level,” he said. “I also learned so much about the new techniques we researched and were able to apply to this problem. We each worked individually and then met to discuss what we found and to develop our model, so it was a great teamwork experience. I now feel very confident handling this kind of problem.”

Developing a proficiency working with unstructured data, or “big data,” was one of Chaturvedi’s motivations for returning to school. After earning his undergraduate degree in information technology from Guru Gobind Singh Indraprastha University in India, Chaturvedi had worked for a few years as a software engineer and obtained some experience in descriptive analytics, but he wanted to expand his knowledge of prediction and visualization.

He choose SILS because of the reputations of its faculty members. He was particularly impressed by Associate Professor David Gotz’s expertise in data analytics and visualization and Professor Arcot Rajasekar’s groundbreaking work in policy-oriented, large-scale data management. Chaturvedi also liked that SILS offered the flexibility that allowed him to combine his information science studies with the STOR curriculum. He said pairing the two is giving him an excellent background in both theoretical concepts and practical applications.

Among other experiences at SILS, Chaturvedi has been working as a research assistant for Associate Professor Brad Hemminger and on projects with Dr. Gotz in the Visual Analytics and Communications Lab (VACLab). Gotz and Chaturvedi developed a working visual analytics prototype using a modeling technique similar to the one applied to the Syngenta problem, and have submitted to present their work at the IEEE VIS conference in March 2017.

Chaturvedi is set to graduate in December and plans to look for a position in a company working on problems similar to the one he and his teammates addressed in the INFORMS competition.

“I want to use data to get insights that can help businesses improve their ongoing processes,” he said. “I’d definitely like to work on a data analytics, data engineering, or data science team.”