UNIVERSITY PARK, Pa. — Like many geography students, Xi Liu has a strong interest in mapping and using geographical data to solve problems. So when he saw an opportunity to work for Google on a project that involved geographical data analysis, he wasted no time in applying.
Liu, a doctoral student in geography, was accepted into a highly competitive software engineering internship program in Google's Seattle office during the summer of 2017, and the experience showed him just how integral geographic data are for the industry giant's products and services.
Google Maps and similar products use algorithms, or complex equations, to provide users with estimated driving time or other requests. These can draw on hundreds of factors, ranging from a road's speed limit to weather conditions to topography to traffic.
Liu spent the bulk of his three months in Google's Seattle offices doing "machine learning," or training a computer how to learn what factors would affect upcoming traffic conditions the most in different cities. After identifying what he thought were the most valuable sources of information, Liu had to do testing to see if a computer algorithm could sift through data from myriad sources and pick out salient information. By the end of the internship, Liu had not only created a model to predict future traffic patterns, but he had also developed a way for his model to "communicate" with other Google products and services.
The project, he said, was "very closely related" to the research he does in geography with his adviser, Clio Andris, assistant professor of geography. Liu's dissertation work involves how to better understand the geography of an urban area by analyzing data provided by people voluntarily, such as Yelp or Airbnb reviews.
"We try to retrieve information from those data and then see if that can be used to help us better understand the function of places, who is using them, and how it affects peoples' sense of connection with their surroundings," he said.
Liu said he has always been interested in people, how they travel and how they connect with one another or their geographical surroundings. But his passion is finding patterns in data, which is why he chose geography for his career path. He received a bachelor's degree from Tongji University and master's degree from Peking University in China, before choosing Penn State based on the reputation of the Department of Geography.
One benefit of Penn State that Liu had not planned for was the Integrative Graduate Education and Research Traineeship, known as IGERT. Many universities offer IGERT programs, and Penn State's focuses on big data and social science. Penn State's IGERT program offers a graduate minor in social data analytics, which Liu completed, providing him with skills in coding and mining data from social networks and platforms.
He also found the broad research areas in the Department of Geography to be a benefit. Researchers study an array of topics, from how diseases spread to how people interpret maps to how satellite and social data can be used in disaster response.
"You can have collaborations with faculty from many disciplines in our department, which can give you a better understanding of people and their behavior in space and time," he said. "This also gives you a different view of the research you've already done, so you can have a deeper and more cohesive understanding of your research problem."
Interning at Google provided insight into how an industry leader uses geographical data — and this is helping Liu home in on what career path he wants to pursue after defending his dissertation in 2019.
"It would be awesome to get an academic job, maybe a postdoc. I'm also open to research positions within industry, which give you access to more data," he said. "As long as it's related to research, I'm open to it."