Leading Holistic Education
Dr. May Yuan of UT Dallas on research, innovation, and the future of GIS
Dr. May Yuan is the Ashbel Smith Professor of Geospatial Information Sciences in the School of Economic, Political, and Policy Sciences at the University of Texas (UT) at Dallas as well as a member of USGIF’s Academic Committee.
Your CV includes many diverse research areas. What are some highlights?
My geography degrees have really helped me to appreciate how humans interact with the environment. GIS is a great platform not just for looking at the natural processes and human activities but also the societal impact. That’s why I’ve had opportunities to look at a wide range of topics, from wildfires to severe weather to patterns of life. The core of my research aims at innovations in methodology—how to think through spatial problems and how to look at spatiotemporal behaviors and interactions of different elements across human and ecological systems. My research group assesses current environments, analyzes relationship, interaction, risk, and impact in space and time, and predicts future scenarios. We seek novel ways to enrich GIS methodology from concepts across disciplines. For example, I collaborated with the Naval Research Laboratory for a project funded by the Department of Homeland Security on spatial risk prediction of suicide bombings applying concepts from ecological species prediction models.
What are you researching now?
I’m working to identify the site and situational characteristics of places most likely to have traffic accidents by time of day and day of the week. Currently, the most common approach is doing hotspot analysis to look at places that have frequent traffic accidents. And there are other models that examine street characteristics such as width, length, curvature, and type to determine their influences of traffic accidents.
Instead, we take a machine learning approach. We segment the street and cut every 100 meters, then for every 100 meters cut to every hour. And based on two years of traffic accident data from 2015 to 2016 in the city of Dallas we analyze that the risk for individual space-time units. We apply the concept of “near-repeat” from criminology to seek cascading effects of traffic accidents in which a preceding accident within an hour and 200 meters might contribute to subsequent traffic accidents. While near-repeat in criminology is isotropic, we introduce the concept of directional spatiotemporal correlation to consider that once an accident occurs it may cause a chain accident to happen. Using machine learning we estimate what the spatial risk will be for a given street segment on a day of the week at a given hour.
The project is funded by the National Institute of Standards and Technology as part of their Public Safety Innovation Accelerator Program and is being done in collaboration with civil engineers at Southern Methodist University. The purpose is to come up with better routing decisions for emergency vehicles, so when they respond to 911 calls we aren’t just giving them the fastest route. We also want to ensure the emergency vehicle can reach the destination safely and avoid accidents that will further delay their arrival. In the long run this could also be used for other purposes. Nowadays so many people use Google or Waze to find the shortest, fastest, or most scenic route. With spatial risk modeling in place one could also request the safest route.
In addition to the traffic risk project, my group is using our campus as a laboratory for smart city research. We are building a comprehensive 3D model of our campus with detailed data on building interior structure, use, and capacity. We are developing algorithms and apps for indoor-outdoor navigation, location-based services for events and parking, and ground-level air quality modeling. We recently formed a drone research team and conduct drone surveys to build high-resolution 3D models, measure near ground air pollutants, and detect and construct archaeological relicts.
What are your thoughts on the future of GIS?
GIS is far beyond a system perspective now. It encompasses multiple sources and platforms to seek computational advances in transforming geospatial data to produce knowledge of our world and intelligence to inform decisions. From an education perspective, we need to incorporate more spatial thinking, quantitative and computational methods, and geographic knowledge epistemology into GIS learning. We shouldn’t focus on teaching students how to use a particular software. We should help students think through spatial problems and know how to develop algorithms and solutions to solve problems. This is quite different from a lot of the current GIS curricula that focus on point-and-click exercises.
What is some of the most important advice you share with students?
I think the most important goal of GIS education is to teach students how to conceptualize a spatial problem and how to develop effective solutions to logically address the problem. Hence, they must be able to recognize the domain knowledge relevant to the given problem and have the abilities to effectively acquire the essential domain knowledge. They need to have good skills in data processing and programming—know where to get the data they need and have the ability to write code to implement their solutions to spatial problems. They also need to have experience with approaches, methods, and multiple software and platforms to be resourceful.
GIS and GEOINT are methodology focused. A common mistake for students is they are holding a hammer and looking for nails. Undergraduate students need to have a good general education background, so they have an appreciation of the basic knowledge of those physical and social sciences. Master’s students need to pick a particular domain area and develop it, so they can apply GIS methodology intelligently to solve problems in that particular domain. Ph.D. students cannot just focus on a particular domain, they need to have a much broader base of understanding physical and social problems, innovate new frameworks for spatial problem-solving, and be able to translate one domain problem to take advantage of methodology involving other domains.
How do organizations like USGIF contribute to the profession?
In my role on USGIF’s Academic Committee, we aim to assist with the accreditation review process and to help enrich the GEOINT curriculum. I hope that through the committee’s efforts we will help a lot of USGIF-accredited programs further enhance their strengths and also improve learning elements that might be lacking. Members of the Academic Committee have played key roles in the development of the GEOINT Essential Body of Knowledge, GEOINT professional certification, and GEOINT education discussions. Many alumni from our member institutions work at various government agencies and in industry. USGIF is a key driver for GEOINT workforce development. In addition to formal education, USGIF also hosts many technical and community workshops for GEOINT professionals.
From my position at UT Dallas, USGIF’s annual GEOINT Symposium is beneficial for the academic community to get a sense of the Intelligence Community and also to hear the directions of the National Geospatial-Intelligence Agency (NGA) and the Department of Defense. It helps us make those connections for students who want to pursue intelligence and defense careers. UT Dallas was fortunate to have two former students hired by NGA this summer. Over time, we will be able to build that alumni base like other universities in the D.C. area have.
Headline Image: Dr. May Yuan after a Vapor 55 drone survey using a PhaseOne iXU 150 camera to develop a high-resolution 3D model of Fort Martin Scott in Fredericksburg, Texas.
Posted in: Education, Perspective, USGIF-Accredited Programs Tagged in: 2018 Issue 4, Education, GIS, Training & Certification
Education, Young Professionals
YPG Park Cleanup
Young Professionals Group gives back to the community at Frying Pan Farm Park
Education, Young Professionals
From Handwritten to Handheld
USGIF contributes to the GIS data collecting efforts needed to help rebuild New Orleans