Sarah Canastra Tackles Sustainability through Data Science and Public Policy
For the past three consecutive years, interdisciplinary teams of students from Georgia Tech have finished first at JUMP into STEM, an online competition sponsored by the U.S. Department of Energy (DOE). The competition focuses on creative ideation in the field of building science, which supports the advancement of building technology and broadly includes aspects such as indoor thermal environment, lighting, air quality, and building resource use. This year’s outstanding team includes Sarah Canastra, a second-year student from the H. Milton School of Industrial and Systems Engineering (ISyE). She worked alongside Hunter Hancock and Lucas Kiefer, who are computer science majors.
Their project, “Load Shifting with Smart Water Heaters: Conservation Without the Cold Showers,” won the Grid Interactive Efficient Buildings (GEB) challenge in the first round of the competition. Canastra went on to present their project at the final event, where she became one of four overall winners and received a summer internship at the Oak Ridge National Laboratory (ORNL) in Tennessee.
The objective of the GEB challenge, one of three competition categories, was to create a conceptual design that optimizes the operation of a building by maximizing energy efficiency. Born out of an initial curiosity about the possible applications of using smart water heaters, the team designed an algorithm to predict water use and recommend the best time to heat water in a home. The project proposal also includes a mobile application for tracking user preferences.
“A smart water heater learns when you need hot water, and it's only going to heat it when you need it,” Canastra said. “We came up with a solution of how to shift the load of the energy grid to reduce the demand during peak demand hours.” This reduces the need for “peaker plants,” which are power plants run during hours of peak energy demand; these are typically more expensive and use more inefficient fuels.
The project is aimed at low occupancy households (LOHs), which are one- or two-member households. LOHs make up over half of all U.S. households, and many have older, less efficient water heaters. Because they have longer periods between hot water requests, LOHs are the perfect target population for the load-shifting algorithm.
“Load shifting works better when there's less people, because if you have all these demand response events – a lot of people saying [they] need hot water right now – [the water heaters] still have to keep heating continuously,” Canastra explained. Their algorithm reduces peak energy usage due to water heaters by 46% on average.
The winning project originated in the Fall 2020 Data Science for Policy course (PUBP 3042) taught by Omar Isaac Asensio, assistant professor in the School of Public Policy. Students in the course were placed into multidisciplinary teams, which aligned with the DOE’s objective to encourage the diversity of thought and background in building science by requiring cross-disciplinary engagement as part of idea submissions.
“In total, we had six Georgia Tech teams compete this year with students from public policy, computing, and engineering,” Asensio said. “We knew we could be very competitive again this year, having won the previous two years in a row. In the end, the excellent research creativity of our students and our particular brand of technology development, integrated with rigorous policy analytics and social theories, proved to be a competitive edge.”
Canastra’s team worked incredibly well together, despite the challenge of being in different time zones, and she enjoyed interacting with other students. “It was really cool to make those connections with people in the class, because it’s hard to get to know people with the online format,” she said.
Another exciting aspect of the project was being able to apply skills Canastra learned from a previous course, Data Input and Manipulation (CS 2316), which is part of the core ISyE curriculum. The course teaches coding in Python, specifically focusing on coding packages useful for data science, which helped her to make data visualizations for the project.
“Having some prior experience working with data [ensured that] I wasn't intimidated by looking at the data sets,” Canastra said. She has already chosen analytics and data science as her ISyE concentration and is minoring in computer science.
This summer, Canastra’s internship at ORNL will be in the building science department, where she will be working on the EMPOWER wall project. The EMPOWER wall is a 5-by-8-foot smart wall that functions as a cooling system for a room with the goals of reducing energy use, decreasing peak time energy demand, and utilizing renewable energy, all while maintaining occupant comfort.
Through its thermal storage and active insulation systems, the wall acts as a thermal battery, adjusting its temperature and the amount of heat released or absorbed according to the season. This reduces electricity costs by lowering the use of the heating, ventilation, and air conditioning (HVAC), which is responsible for a large percentage of energy use in buildings. Canastra will be involved with curating and analyzing data from the wall to evaluate energy and peak demand savings.
Canastra’s experience in her public policy class also led her to Georgia Tech’s Data Science and Policy Lab, directed by Asensio. Currently, she is involved in a sustainable plastics project funded by the National Science Foundation. As part of the beginning phase of the project, she is researching potential data sources, and working on a literature review of different plastics policies and their effectiveness.
Originally, Canastra began her studies at Georgia Tech as a pre-health chemistry major, but she quickly realized that pre-health was not the right path for her. When she started researching different majors and came across ISyE, she was intrigued by the combination of math, computer science, and business concepts in the curriculum.
“One of the things that drew me to [industrial engineering] was how many different types of classes there were,” Canastra said. She is interested in the applications of ISyE to public policy and sustainability, such as making systems more efficient and better for the environment.
These incredible experiences – winning JUMP into STEM, joining the Data Science and Policy Lab, and her upcoming internship at ORNL – all came out of her decision to take an extra class. Canastra already had a prior interest in public policy, and when she discovered Data Science for Policy, she knew it would be the perfect combination of her passions. Balancing a busy course load can be challenging, but she encourages other students to explore classes outside of their majors as well.
PUBP 3042 Data Science for Policy meets Georgia Tech’s general education requirements for Area E: Social Sciences, making it accessible to undergraduate students broadly across the Institute. In the next offering of the course, PUBP 3042 will also open available seats as part of Georgia Tech’s Honors program.
- Workflow Status: Published
- Created By: goberst3
- Created: 05/13/2021
- Modified By: Shelley Wunder-Smith
- Modified: 05/20/2021