You are here

Superpowered GIS

October 6, 2017

Eric Shook transitioned from computer science to geography in two days. He was working at the University of Iowa on a grid computing project related to the Large Hadron Collider, when his supervisor asked if he’d like to pursue a PhD in geography at the University of Illinois.

“I talked to my wife that night, and I told him we’d drive to the campus for the weekend and check it out,” says Shook. “He said, ‘no, I need your decision by tomorrow.’ So we decided to give it a go.”

Today, Shook continues to take a computational approach to geographic information systems (GIS). He brings this unique outlook to his work at the University of Minnesota, where he has served as an assistant professor since 2016.

Given his background, Shook often acts as a liaison between computational systems and GIS. “I sometimes introduce myself as a ‘supercomputerist,’ because my expertise is really in trying to facilitate getting science done on supercomputers.”

Currently, he’s working on a programming language called FOREST (For Expressing Spatio-Temporal computing). FOREST is written in Python, a programming language that’s easy to learn.

“The whole idea behind the language is to make big geospatial data processing easier to do in research and teaching,” says Shook. “One of the motivations for creating it is that I can introduce it in my own classes, to both graduate students and undergrads.”

Shook also spends a lot of time observing Twitter data and trends. GPS locations are embedded within some people’s tweets, which can reveal how people are responding to major events. This work is part of the Socio-Environmental Data Explorer, a web-based tool designed to advance risk theory.

“We collect more than 1 million tweets a day, and we’ve been collecting data for two years, so we’ve accumulated over 2 billion tweets so far.” This real-time data enables scientists and decision makers to examine public perception of risks.

“There are some events where people think it’s riskier than it actually is—or, they assume there isn’t much risk when there really is,” says Shook. Case in point: Flint, Michigan, put a spotlight on the dangers of lead-contaminated water, which amplified the risk. But evidence shows that lead in paint is an even bigger risk in the U.S., though that issue doesn’t get as much attention in the media.

Ultimately, all of this data can improve how society responds to risks. It can also reveal how those risks reshape communities in the long term. Are they more resilient or less resilient? Do they grow so accustomed that they downplay the dangers? These are the kinds of questions that CLA faculty are trying to answer, helping to create a more powerful story.

“One thing that liberal arts does better than some of the other sciences is adding narratives,” says Shook. “And narratives are what weave the liberal arts and big data together.”