Spatial Dependence in Data
From a bird’s-eye-view, Professor Veronica Berrocal’s research focuses on developing statistical models that extract information from data collected over space and time. “My research is centered around developing statistical models that can describe the interdependence and correlation among observations, how that dependence varies in space (and time), and how to account for it while addressing certain questions,” says Berrocal. “For example, what is the level of air pollution where we don’t have monitors and how can we infer upon how poverty in certain neighborhoods changes over time from coarse, temporally aggregated data?”

Post-Processing Outputs from Geophysical Models
In various Earth Sciences disciplines, scientists use computer models to understand and predict spatio-temporal processes (such as climate or air quality models). These models often present issues of calibration, showing (spatial) discrepancies with observed data. Professor Berrocal’s work in this arena is focused mainly on developing statistical methods to calibrate the output of geophysical models while also addressing the difference in spatial resolution between the observations and the model output. She also works to combine observational data with computer model outputs to better estimate the physical spatio-temporal process of interest.

Neighborhoods Characteristics and Health
“Characteristics of the environment where people live influence people’s lifestyle choices and thus can contribute to the adoption and maintenance of health promoting behaviors,” says Professor Berrocal. On one front, she is collaborating with researchers at Drexel University to develop statistical approaches that help identify the spatial and temporal scale of relevance for the effects of the built environment on health. For example, she’s exploring how the spatial distribution of fast food restaurants around schools affect the risk of obesity in school children. She is also working with researchers from the University of Michigan to develop approaches that leverage social media to obtain spatially detailed, actionable measures of a neighborhood’s health behaviors.