In ‘Plus’ programs, Students Dive Into Pressing Issues

A lot of people gathered with posters
A showcase of work stemmed from Duke’s summer “plus” programs, including  Data+, Code+, Computer Science+, Applied Ethics + and Climate+ .

Last month, historic, intense rainfall by Tropical Depression Chantal triggered flash flooding in Durham and the surrounding Triangle region.

The storm left streets underwater, prompting boat rescues in Durham County after the rainfall caused the banks of the Eno River to overflow.

While the Triangle did not report any fatalities like the death toll in Kerr County, Texas, including the tragedy at Camp Mystic, the flash flooding nonetheless gave rise to safety concerns among Triangle residents.

So, it was fitting that in late July, Duke undergraduate Suhaib Mansour unveiled a project he had spent the summer working on, titled, “Forecasting heavy rainfall in the Southeastern U.S. using deep learning.” His project was one more than 50 shared at a showcase of work stemming from Duke’s summer “plus” programs, including  Data+, Code+, Computer Science+, Applied Ethics + and Climate+ .

Solutions to real-world problems

Students, faculty and staff gathered at Gross Hall to discuss the projects they were working on during the 10-week summer program, which pairs undergraduate students with experts to explore solutions to real-world problems related to health, the environment and society.

Mansour, a native of Orlando, Florida, worked in collaboration with students from Carnegie Mellon University in Pittsburgh and Kaust Academy in Thuwal, Saudi Arabia. The trio studied flash flooding forecasts in the southeastern United States.  

Their work was displayed on an oversized poster among other ongoing projects. The other projects presented included: “Understanding the Learner Experience from Submissions and Error Patterns,” by Amir Aref, a rising senior at Berea College in Berea, Kentucky, that explored concepts in programming or curriculum designs that would cause students to struggle when it comes to learning computer programming. 

Another project, “Control-Flow Detection Using KLEE,” by Duke undergrads Yuxiang Liu and Jameson DiPalma, relied on static analysis to identify computer security violations and bugs.

“It’s huge on cryptographic libraries, where it’s very important that secrets stay secret,” said DiPalma, a rising senior from Rochester, New York. 

From local to global

Some of the projects were local, such as the “Duke University Greenhouse Gas (GHG) Data System Update,” and others were global, including “Mapping High Stakes Coastal Zones: An Interactive Dashboard for Improved Coastal Decision-Making,” which explored coastal regions across the world that are most vulnerable to climate change.

One project explored vulnerable coastal regions across the world.

“We created an interactive dashboard that takes climate-related data from different sources and puts them together with emphasis and focus on the coastal areas,” said one of the project members, Leonard Eshun, a Duke graduate student and native of Ghana.

“Researchers, policy makers, and NGOs can look at the dashboard and have an overview of the climate effects around the globe, zoom in to get detailed information at a regional level, and decide where and how to effectively focus their efforts,” added Eshun.

Meanwhile, Mansour stood alongside the poster that displayed graphics on his work. “So, for our project, we built a deep learning model to forecast heavy rainfall events in the southeastern United States,” he said. The model created by Mansour and his colleagues is a 40-minute forecast with about 76 percent accuracy.

Mansour said the model isn’t ready yet for prime time and real-world applications, but he envisions its use in the future as a “specialized tool.”

“I think if people realize how similar we are as a community ... then people would be friendlier to each other. [The goal] is to make this data more accessible to the public.”

Joie Jacobs, ‘Plus’ Program Participant

“If we have a two, three or four hours forecast, that’s enough time for people to evacuate,” he said.

Joie Jacobs, from Mount Laurel, New Jersey, will begin her sophomore year at North Carolina Central University this month. She spent the summer working, with Duke students Alexa Fahrer and Tim Le, on “Visualizing polarization in American public opinion.” The project relied on 2024 American national election survey data.

“We’re looking to see how the American public feels about polarizing topics such as abortion, health care, immigration and gun policy, for example,” Jacobs said.

“I would say every human has the same basic needs and wants psychologically, intuitively and in nature,” Jacobs said when asked if there were any hopeful takeaway points with the project.

“We are not as polarized on topics that meet people’s basic needs, like health care, for example, and Social Security,” she added. “I think if people realize how similar we are as a community, especially considering the United States as a melting pot, then people would be friendlier to each other. [The goal] is to make this data more accessible to the public.”