Centers for Disease Control and Prevention Fellow uses data science to help create activity-friendly communities Meet Graycie Soto

Movement has always been a passion for Graycie Soto. She grew up playing sports, and as a young adult developed an interest in jogging, walking, and biking through natural and urban environments. Soto turned this passion into an exercise science degree from Belmont University, before transitioning into the public health field. She received a Master of Public Health from George Washington University and sought to help others become more physically active on a wider scale.

As Soto moved and travelled around the U.S., she became interested in structural impediments to physical activity.

“I saw how different communities face systematic barriers to physical activity, such as access to safe places to be physically active, that led to disparities in health,” says Soto.

So, when Soto came across the Center for Disease Control and Prevention (CDC) Research Participation posting in the Oak Ridge Institute for Science and Education (ORISE) newsletter, the opportunity was too good to pass up.

“I applied because this program was right up my alley to gain experience in analyzing data, use the findings to inform public health action for communities to increase access to physical activity, and be a part of a diverse team of public health professionals,” says Soto.

Soto applied and was accepted for a fellowship with the Physical Activity and Health Branch in the Division of Nutrition, Physical Activity and Obesity (DNPAO) within the National Center for Chronic Disease Prevention and Health Promotion (NCCDPHP) at the Centers for Disease Control and Prevention (CDC).

The CDC Research Participation Programs are educational and training programs designed to provide students, recent graduates and university faculty opportunities to participate in project-specific CDC research, current public health research and developmental activities.

Now, four years into her fellowship, Soto’s typical day is usually spent somewhere in the research process either developing research questions, writing code to conduct an analysis, or writing and editing. While her research is multifaceted, the overall purpose is to increase opportunities for safe and accessible physical activity for all people, regardless of age, race, ability or geographic location.

Soto has been involved in three main areas of research during her time in DNPAO’s Physical Activity and Health Branch while working with her mentor, Dr. Miriam Van Dyke.

Her first research project looked at people’s perception of traffic as a barrier to walking safely near their homes.

“Results from this study can help inform communities to consider speed reduction or improved sidewalks as a strategy to develop safe access to physical activity,” she says.

Her second research project involves using modernized methods for physical activity research. One of these methods includes using anonymous location-based big data from smartphone applications. The purpose of this study was to find whether data from location-based services from cell phones is valid for measuring physical activity in communities in a timely way as compared to public health survey systems with longer lead times. The study, co-authored by Soto and Bryant Webber, finds that smartphone data may complement traditional public health data for community-level physical activity patterns.  

In the future, these location-based data may “provide insight into which parts of a community most need built environment interventions to make physical activity easier and more accessible for everyone,” Soto says.

Soto’s third project involves using automated image analysis of street-view images to gather data on the built environment. She notes how this practice may have the potential to “generate an important data source for public health officials and policymakers to consider when addressing and improving the built environment.” While there’s still much to learn, these data could also be used to identify sidewalk locations, essential for creating activity-friendly communities that provide connections to everyday destinations.

Last year, Soto and two of her colleagues had the unique opportunity to participate in the Data Science Upskilling (DSU) program at CDC. This 10-month-long team and project-based data science training program allowed Soto to work on a project called, “Automating Extraction of Sidewalk Networks from Street-Level Images.” Soto and her team used machine learning algorithms in Python to train models to identify the presence of sidewalks in street-view images using computer vision. “Basically,” she says, “we wanted to learn how to train a machine to predict if there is a sidewalk present in a street-view image.” Soto found this opportunity incredibly rewarding.

 “The skills I took away from this experience are unrivaled. I learned data science skills that I never imagined I’d have – such as knowing how to build a machine learning model and understanding how it works!”

Overall, Soto’s impression of the ORISE fellowship experience is that it is “invaluable.”

“It has provided tools to take with me as I grow my career. It has opened up doors for me professionally and enhanced my skills,” she explains. “Ultimately, being an ORISE fellow gives a feeling of accomplishment and knowing that the work we are doing is truly making a difference.”

Soto hopes to continue supporting her work in DNPAO’s Physical Activity and Health Branch even after her fellowship concludes.

Soto wants to emphasize how “improving physical activity is a whole systems approach with drivers from behaviors to community and policy.” While this is a “complex process,” she and her team have provided data and research that could make communities better places to live. 

The CDC Research Participation Program is managed by the Oak Ridge Institute for Science and Education (ORISE) under an agreement between CDC and the U.S. Department of Energy (DOE). ORISE is managed for DOE by ORAU.