About Angela Chen
Inspired by a family friend who almost lost their life during a fire, Angela developed a machine learning model that studied wildfire risk in California, taking into account nine different environmental variables. She hopes her model will better predict burn areas and regional wildfire risks in the coming years.
Developing Wildfire Risk and Burn Area Prediction Models Using Comprehensive Environmental Variables for California With Machine LearningView Poster
Angela Chen, 17, of Cary, developed models for predicting California wildfire risk and burn areas based on nine environmental factors for her Regeneron Science Talent Search environmental science project. After a close family friend was nearly killed in a wildfire five years ago, Angela developed a machine learning model to quantify the effect of drought on overall burn area and published her findings in the International Journal of Wildland Fire as sole author. Angela’s project expands on that research to add eight more variables, such as wind speed and soil moisture, and quantifies their impact on fire risk and burn area. Drawing from 36 years of publicly available data, she developed models to calculate possible burn areas and regional wildfire risks based on those nine environmental variables. With a prediction rate of over 90% accuracy, Angela hopes her research can improve forecasting ahead of the state’s future fire seasons.
At North Carolina School of Science and Mathematics in Durham, Angela is co-president of the speech and debate club. She previously co-founded a nonprofit that arranges tutoring and tournaments for chess players from marginalized groups. Her parents are Chunling Tang and Dong Chen.
Beyond the Project
Angela began studying wildfires in seventh grade and published early findings in the International Journal of Wildland Fire as sole author.
FUN FACTS: A team captain of her school’s environmental competition club, Angela loves frogs and has been known to sport frog phone cases, wear frog jewelry and pass out frog stickers.