For two months, three graduate students at the University of Maryland University College (UMUC) brainstormed on a project that used climate data to foresee outbreaks of Lyme disease. For their effort, they beat out 210 other university teams for the top spot in the prestigious Watson Analytics Global Competition.
The UMUC students—Tracy Carr, Elizabeth Handy, and Abebaw “Abel” Zeleke—share the top honors with a team from Canada’s Dalhousie University.
“This is a very impressive achievement,” said Elena Gortcheva, chair of UMUC’s Master of Science in Data Analytics program and the team’s mentor. “These were three working professionals all in different companies, all taking classes in the online master’s degree program, so they were people without a lot of extra time. They worked very hard on this project.
“For IBM, the most important [factors] were the originality of the topic, the application, and how the team used the different features of the software,” Gortcheva explained. She described the winning students as “original thinkers.”
All three students will complete an M.S. in Data Analytics in August.
IBM’s annual competition seeks innovative analytics solutions from students using the company’s Watson Analytics platform. Each three-member student team must employ existing public databases in identifying a problem and its solution. The 2017 competition focused on the theme of global environmental issues. Carr, Handy, and Zeleke won for their report, “Climate and Landscape Change Effects on Lyme Disease.”
“Finding a focus was the biggest challenge of the competition,” said Handy, the director of marketing at Maryland University of Integrative Health. “Everything we wanted to dive into was too broad or didn’t have granular enough information.”
In fact, the winning entry was not what the team originally had planned.
“We started off with a different idea, the impact of freshwater availability, which was more global. But we had trouble finding data,” said Zeleke, who works as the technical lead for General Motor Corporation’s Enterprise Application Integration Team in Atlanta. “Beth [Handy] brought up the idea of Lyme disease.”
The incidence of Lyme disease, which is transmitted through ticks, has doubled since 1991, making it the most common vector-borne disease in the United States. The team decided to see how climate change may factor into the growing transmission rate.
“I see that others in the competition wrote about sea levels rising, global warming. These are problems that are not new,” said Gortcheva. “But Lyme disease impacts health. To look at that showed innovation.” The professor said the project also turned research into something with real-life application, a goal that defines the data analytics program at UMUC.
At the end of 2016, the team decided to participate in the competition and then, as Handy put it, “like all good graduate students we took the holidays off.” When they regrouped, they began work on the freshwater project they had envisioned, scrambling toward an April deadline.
Zeleke said the “darkest moment” came in late February—several weeks into the project—when the team decided it didn’t have enough material to pursue its freshwater idea and it would have to find a new focus. When the students settled on Lyme disease, their time on the project stepped up dramatically.
“In the month of March, I did nothing but go to work, come home, and work on the project,” Handy said.
Even after switching the focus, finding relevant databases was difficult. The students looked at how temperature and weather indicators, as well landscape and habitat changes, may affect the disease upswing. Among the team’s surprising choices was the use of building-permit datasets, at Carr’s suggestion, to correlate whether construction has dislocated populations of deer, which are carriers of the tick that causes the disease.
When deer lose habitat and are pushed into areas that are already dense with other deer, their contact with ticks that transmit Lyme disease steps up.
Carr, a data analyst at Wells-Fargo in Charlotte, North Carolina, said that her team addressed the subject from a public health perspective. Since prevention efforts and early diagnosis of Lyme disease can avert the most dangerous health effects of the illness, the ability to predict where it might spread is valuable.
“We looked at building-permit datasets nationally, and we concentrated on four states in New England with the highest growing incidence rate [of Lyme disease],” said Handy. “Our working theory, which is not scientifically proven, is that as tick and deer populations get more concentrated and weather conditions are ideal for tick breeding, the disease spreads. Those deer travel into populated areas because there isn’t enough space for them where they are.”
In the first round of competition, each team presented its solutions virtually via a written report and a video. The winners were announced on May 31.
In mid-July, the UMUC and Dalhousie University teams will travel to Malaysia to present their solutions at the 5th International Big Data and Analytics Educational Conference. That gathering on Langkawi Island will mark the first time that members of the UMUC team have met one another in person.
Throughout the project, the students worked together in a virtual setting, using a password-protected group folder. Yelena Bytenskaya, an adjunct faculty member who works at GP Strategies Corporation, an e-learning company, coached them on how to use the Watson Analytics software. The three graduate students collaborated on the research then delegated duties for the final presentation.
“Tracy took the lead on the video presentation. I wrote the paper. Abel was the lead on the data piece of it,” Handy said.
The three said they chose UMUC’s data analytics program because, as working professionals, they needed the flexibility offered by an online program. Carr, too, noted that her parents and brother are University of Maryland graduates.
For more information about the team’s winning project see their YouTube video presentation, “Climate and Landscape Change Effects on Lyme Disease.”