by Bryan Lindenberger
Computer technologies have become vital in nearly every industry from home building to healthcare. Geosciences certainly has felt the impact of computer science with geographic information systems (GIS) marrying drones to advanced software packages to create 3D maps of our world.
Youngho Lee and Yubin Lee – no relation - are using computer science to advance geosciences in an entirely new way. As visiting scholars to the University of West Georgia from Kyung Hee University (KHU) in Seoul, South Korea, they work with advanced machine learning systems to find long-term trends in geographic research.
“The Association of American Geographers (AAG) provided us with between six and seven thousand abstracts of projects presented at their conferences over the past 20 years,” said Dr. Jeong Seong, professor of geography in the Department of Geosciences at the College of Science and Mathematics.
“We are attempting to visualize the keywords in these abstracts to better understand what topics are presented and what trends have developed over time,” Seong said.
Funded by a Korean government grant secured by Dr. Hwang of KHU, the project is called High Potential Individuals Global Training Program. Awarded approximately 50 thousand dollars, Seong serves as principal investigator on the U.S. side.
Much of Seong’s role consists of guiding visiting scholars Youngho Lee and Yubin Lee in their research. Their task is to create visual representations of trending topics in geography research.
“We use the abstracts provided by the AAG as our source,” Seong said. “The resulting visualizations include clouds, heat maps and information graphs.”
Even if you are unfamiliar with these terms, you likely have come across their uses in daily life.
Word clouds, for instance, are popular on the Internet. They depict clusters of words drawn from a website, social media page or other data source. The most frequently used words appear larger and bolder in the cloud, denoting their use and prominence.
Similarly, information graphs – or igraphs – represent how keywords connect to each other and cluster around topics. At the UWG Athletics page, the topic “sports” would appear most prominently and connect to keywords or subtopics such as “football,” “golf” and “excellence.”
“Through big data analysis of keywords, titles and other data minded from the abstracts, we can create a visual representation of the abstracts in sum,” Seong said. “This will help us better understand where research in geography has been and where it is going.”
Of course, no human could accurately analyze such large quantities of data. Even if they could, human biases would creep in and sully the data.
That is where Dr. Ana Stanescu, assistant professor in the Department of Computer Sciences, plays a key role as co-principal investigator.
An expert in machine learning and intelligent systems, Stanescu knows the means, methods and materials required in this type of big data analysis. More importantly, she provides the guidance and skill in understanding how to use intelligent systems to extract the best, scientifically sound data.
“The first step is translating the abstracts from PDF into plain text,” Stanescu said. “Then we can extract the data and begin building networks.”
The software used in this process includes R for data mining and visualization as well as Python and the Cytoscape suite for visualizing complex networks as igraphs.
“This type of meta-analysis is very new,” Stanescu said. “No one has done this before for AAG. We are taking part in a knowledge evolution that will be transferable to other disciplines.”
For their part, the visiting scholars are learning a great deal at UWG. While both are graduate students of geography, the computer science aspect is new to each of them, as is living in Carrollton.
“In Korea, most students do not usually say anything or ask questions,” Youngho said. He noted the open curiosity he saw in the American classroom. “It is interesting to me, and I like that the students here ask questions during class.”
Youngho also greatly enjoys the food at East Commons, a fact with which Yubin immediately agreed.
“I love East Commons!” she exclaimed.
Yubin also said that she enjoys the American atmosphere that allows for more free time and personal exploration outside of classes and labs.
“I like the nine-to-five system,” she said. “In Korea, we usually work to 7 or 8 p.m.”
Both students are currently taking classes that include Intelligent Systems, their first foray into computer science. The project, and their time in the Carrollton, extends until June 30, 2020.Posted on