Select Publications

Using latent class analysis of 5,280 Kansas PRAMS participants, this study identifies five distinct antenatal stress profiles and links them to adverse birth outcomes including low birth weight, small for gestational age, and preterm birth.

This study analyzes U.S. nonfatal self-inflicted firearm injuries using clinical narratives and machine learning to uncover distinct racial and ethnic patterns and improve injury prevention strategies.

This study investigates the lasting emotional and physical symptoms experienced by survivors of intimate partner violence (IPV) in the United States. Analyzing national survey data from 2009 to 2023, the researchers examined whether sustaining a serious physical injury (such as broken bones, gunshot wounds, internal injuries, or loss of consciousness) is associated with subsequent health symptoms. The findings reveal that IPV survivors who sustain serious injuries have significantly higher odds of experiencing both emotional symptoms (such as feeling vulnerable, unsafe, or distrustful) and physical symptoms (including headaches, chronic fatigue, and muscle tension). This highlights the compounding burden of physical injury on overall well-being and emphasizes the critical need for integrated medical, legal, and mental health interventions to support survivors.

This project applies structural topic modeling (STM) and natural language processing (NLP) to strengthen structural mitigation in capital defense.

This study uses artificial intelligence and computer vision to analyze street-level imagery from Google Street View across 141 census block groups in Los Angeles, California. By measuring features of the built and natural environment, the research team constructed neighborhood indices for natural surveillance, lighting, land-use mix, and greenness. The findings show that higher quality built environments, specifically better lighting, visibility (natural surveillance), and green space, are associated with lower rates of neighborhood child neglect, even when controlling for area-level deprivation and poverty. This research demonstrates how AI street-level audits can provide actionable data for city planners, social workers, and policymakers to design community-based prevention strategies that protect vulnerable children.

This study explores the associations between developed open space, tree canopy, blue space, and land cover diversity and suicide risk across Chicago neighborhoods. Using Bayesian spatial Poisson models at the Census Block Group level, the research reveals a land cover paradox: while developed open spaces and blue spaces are associated with lower suicide rates overall, the protective effects are context-dependent and are not uniformly experienced across urban neighborhoods. The findings emphasize that suicide prevention efforts should incorporate neighborhood-level environmental factors and structural disadvantages rather than focusing solely on individual risk profiles.

This study applies machine learning and natural language processing to analyze child fatality reports in Pennsylvania from 2016 to 2023. Unstructured text narratives from child death investigations were combined with census data to identify recurring themes and environmental factors surrounding child neglect and traumatic injuries. Using structured topic modeling, the researchers identified 11 distinct circumstances of death, including substance misuse, sleep-related deaths, firearms, and supervisory neglect. The analysis showed that counties with higher poverty levels and larger non-White populations experienced a higher prevalence of firearm-related and severe traumatic injuries. The findings demonstrate how combining clinical narratives with machine learning can strengthen fatality surveillance and help construct more effective, targeted prevention programs.

This study examined how pedestrian-level green space relates to firearm-related and natural deaths in Cook County, Illinois, using a matched case-control design. Various greenness measures, including street-view green imagery, tree canopy coverage, and green space along walkable streets, were inversely associated with firearm homicide risk. Protective effects of tree canopy and localized greenery were stronger in less deprived neighborhoods, while neighborhood racial composition influenced the relationship between park proximity and homicide risk. Overall, street-level green space is linked to lower homicide odds and may offer greater protection in vulnerable communities.

This study describes the development and validation of a machine learning model (random forest) to predict the risk of firearm-related injuries across Chicago neighborhoods. By combining multi-source data, including local crime reports, socioeconomic vulnerabilities, and environmental exposures, the model maps out areas with the highest predicted risk. The model’s predictions were validated against actual administrative records to verify its spatial accuracy. This predictive modeling framework shows how machine learning can help public health officials and community organizations anticipate high-risk spatial hot spots and deploy community-based interventions more effectively.

This study examined the spatial overlap of child welfare and juvenile justice involvement using multivariate Bayesian spatial modeling. Results show a significant positive association between juvenile justice and child maltreatment referrals across neighborhoods. After adjusting for residential instability, immigrant concentration, and residential segregation, children in the least deprived 20% of neighborhoods were 95.2% and 55.5% less likely to be referred to child welfare or justice systems, respectively, compared to those in the most deprived 20%.

This research presentation examines the interplay between the built, natural, and social environments in predicting the risk of child abuse and neglect (CAN). By utilizing machine learning algorithms, the study categorizes environmental determinants of child well-being into key dimensions including urban design, environmental contaminants, and neighborhood socioeconomic conditions. The study demonstrates that characteristics of the physical and natural environment, such as neighborhood walkability, lighting, and greenness, interact with social vulnerabilities to shape family stress and safety. This work highlights the value of using computational models to translate physical streetscape characteristics into policy-relevant indicators for child maltreatment prevention.

This study investigates how individual case characteristics and county-level social vulnerabilities influence the decisions made by child protective services (CPS) investigators when recommending services for families. Analyzing over 12,000 cases across New Mexico, the study looks at factors such as child age, abuse type, prior history, and county-level deprivation (e.g., housing and transportation deficits). The results show that county-level characteristics significantly influence whether a family receives support services. Even when families present with similar risk levels, those living in counties with higher structural vulnerabilities are subject to different service disposition outcomes. The findings highlight the need for standardized decision-making tools to ensure equitable treatment and service distribution across different geographic regions.

This study examines the geographic and temporal patterns of child abuse and neglect referrals across New Mexico census tracts over a 9-year period. By utilizing Bayesian spatial models, the researchers identified areas with persistently high risk and examined how social and housing vulnerabilities affect maltreatment rates. The results demonstrate that neighborhood deprivation, housing insecurity, and residential instability are strong predictors of substantiated child abuse and neglect. The study highlights the utility of using space-time models to pinpoint communities that require enhanced support services, enabling child welfare agencies to allocate resources more equitably and proactively.

This study evaluates the exposure of school children to gun violence along their walkable routes to and from school in Englewood, Chicago. Using acoustic sensor detection technology (ShotSpotter) alongside realistic network routing models, the research quantifies both direct and indirect gunfire exposure during daily school commutes. The findings show that children walking to school face frequent and geographically concentrated gunfire exposure, which varies significantly depending on specific walking routes and neighborhood social vulnerability. This work demonstrates the power of integrating spatial routing algorithms with acoustic gunshot data to guide school safety policies, active transit planning, and targeted community policing.