Story URL: http://news.medill.northwestern.edu/chicago/news.aspx?id=200403
Story Retrieval Date: 12/21/2014 8:05:20 PM CST
Merging the worlds of ecology and computer science meant a trip to the Mpala Research Center in Kenya for Chicago graduate students.
Students enrolled in field computational ecology at the University of Illinois at Chicago made the trip to Kenya in January. In Africa, they conducted experiments, used computer technology to gather wildlife data, and took 3D images and video footage to create Mpala as a virtual space back home.
The UIC's Electronic Visualization Laboratory and Laboratory for Computational Population Biology worked on the initiative with scientists from Princeton University.
LCPB Director Tanya Berger-Wolf combined her background in data mining, which focuses on finding patterns in information, with EVL professor Jason Leigh’s expertise in data visualization. And while the faculty organized the course, it was the students in her ecology course who drove the projects, Berger-Wolf said.
Among these projects were the study of ants in acacia trees, ant navigation and a “Virtual Mpala” project.
For the acacia tree project, the students are working on a mobile app that uses GPS to help ecologists plot individual trees. For the ant navigation research, they have developed an algorithm so that an ecologist does not have to watch hundreds of hours of video to determine which direction ants are going.
For the “Virtual Mpala” project, students documented everything that went on in the research center through 3D images and movies. The images would then be used to simulate field experience before future students go on the trip.
This year’s trip expanded on experiences from a 2010 trip. “This time, we got funding, we got preparation, we got some background lectures for the students,” Berger-Wolf said.
The collaboration between the two universities developed over the last seven years, when Berger-Wolf teamed up with Dan Rubenstein, a behavioral ecologist at Princeton University.
In 2010, without much preparation, Berger-Wolf said UIC “hijacked” Princeton’s existing course in “Tropical Ecology.”
“We thought, ‘Great! We’ll just add computer science to it!’” she said.
This January, the students worked in an intriguing, but “not surprising” collaboration with the Princeton ecologists, Berger-Wolf said. For some UIC students, it was nice just to get outside. “We usually work in the lab where, as you can see, we don’t have a lot of fresh air,” said Victor Mateevitsi, a PhD candidate at EVL.
“It was nice, but you got very tired. Going out in to the field and actually collecting data, you realize that it’s very difficult,” he said.
When working in tandem with the ecologists, “You value when someone gives you data and you understand that it’s difficult for them to get it,” he said.
The trip had its snags, said Berger-Wolf. For example, when the computer scientists set up camera traps to document animals in the area, they ended up with 7,000 photos per day.
The motion-sensor cameras ended up capturing more than what they wanted - because of the grass.
“Twenty-five percent of [the photos] was grass moving in the wind,” Berger-Wolf said. “Some poor ecologist, after looking at two days worth of data—14,000 pictures—got up from her chair and, in a sentence about two-thirds of which was expletives, said that she ‘is done’.”
One of the students is working on a way of eliminating all false positives, such as grass. The program would also tell the difference between herbivores and predators, and whether the animal is grazing or lying down, based on the position of its head.
Working in the field gave the UIC students a better sense of where data comes from.
“Without going there, there’s no way for me to understand the problem as I do now,” said Khairi Reda, an EVL PhD student.
Students said the two universities complimented each other well with the balance of ecology and computer science.
“I personally think that this is the frontier of collaboration, is to work together as a team,” Reda said.