About

Kathleen Stewart is a Professor and the Director of the Center for Geospatial Information Science (CGIS) at the University of Maryland. Her work focuses on geospatial sciences, with particular emphasis on remote sensing, human dimensions of global change, and landscape-scale processes. As a professor, she contributes to the advancement of geospatial analytics and earth observation, integrating these technologies to address global environmental challenges.

Research topics

• Computer Science
• Statistics
• Environmental health
• Medicine
• Political Science
• Geography
• Remote sensing
• Mathematics
• Atmospheric sciences
• Demography
• Psychology
• World Wide Web
• Physics
• Environmental engineering
• Business
• Finance
• Public relations
• Telecommunications
• Environmental science
• Chemistry
• Meteorology
• Internet privacy

Selected publications

• Detection of non-recurrent traffic congestion over complex road networks using distributed computing and an unsupervised deep learning framework

  SSRN Electronic Journal · 2026-01-01

  preprintOpen accessSenior author
  PublisherDOI

• A Place-based Network Approach to Wildlife Poaching: Integrating Network Analysis and Crime Scripting

  2026-04-02

  article

  Wildlife poaching poses a serious threat to global biodiversity and species survival, yet effective prevention remains challenging due to its complex spatial and behavioral dynamics. Addressing this multifaceted problem requires interdisciplinary approaches that integrate crime science with geospatial analysis. Rooted in environmental criminology, crime scripting provides a framework for understanding the sequential processes of crime commission, while network analysis offers a place-based perspective for examining how illegal activities are spatially organized. Using poaching in Pù Mát National Park, Vietnam, as a case study, this study constructed a spatially explicit place network dataset structured by four crime-script stages: Preparation, Pre-activity, Activity, and Post-activity. Spatial social network analysis (SSNA) metrics were used to examine network structure and spatial characteristics, including structural density, spatial compactness, directional entropy, and movement distances. Community structure was detected using the Louvain algorithm, and centrality measures identified influential hubs, bottlenecks, and persistent places and movements relevant for intervention prioritization. Results showed that poaching place networks were structurally sparse yet spatially diverse, exhibiting strong community structure and stage-specific spatial patterns. Although simple in structure, poaching activities could be strategically organized through locally structured place-groups, spatial partitioning, brokerage places and movements that facilitate coordination across stages. By integrating crime scripting with SSNA for the first time, this study advanced spatial understanding of wildlife crime and provided actionable insights for place-based intervention. The proposed analyses package is transferable to other conservation crimes contexts to support targeted, stage-specific prevention strategies beyond isolated hotspot interventions.

  PublisherDOI

• Insights on Late-Stage COVID-19 Pandemic Recovery From a 21-Country Online Survey

  International Journal of Public Health · 2025-03-28

  articleOpen access

  Objectives: The widespread impact of the COVID-19 pandemic on health systems, economies, and societies globally requires comprehensive data to guide effective recovery efforts. Online surveys have become crucial for rapid and extensive data collection. The Pandemic Response Survey (PRS), utilizing the Facebook Active User Base (FAUB), assessed the pandemic's population-level impacts across 21 countries, gathering information on healthcare, vaccine confidence, trust, and economic and educational indicators. Methods: Conducted from March to May 2023, the PRS, translated into 15 languages, used the FAUB for gender-stratified random sampling of adults 18 years and older. The survey collected responses from 621,000 individuals, achieving a completion rate of 43%. Non-response and inverse propensity score weights were applied to calibrate the data to known demographic totals, enhancing the survey's generalizability. Results: The PRS findings reveal disparities in life satisfaction, food security, delayed healthcare, vaccine confidence, and trust across countries. Life satisfaction was reported as high by 70%-80% of respondents in Egypt, Nigeria, Colombia, and Mexico, while only 20%-30% of respondents in Indonesia, Turkiye, and Viet Nam reported the same. Approximately 50% of respondents in Nigeria, South Africa, and Colombia experienced food insecurity, in contrast to less than 10% in Italy, Japan, and Germany. In Germany, 44% of respondents expressed high vaccine confidence compared to 10.6% in South Africa. Over half of respondents in Indonesia (52.4%) reported that their child was up to date on routine immunisations. Conclusion: The PRS demonstrates the effectiveness of online surveys in capturing actionable data during a global health crisis. The findings underscore the importance of targeted interventions and policy decisions to address the multifaceted challenges of pandemic recovery. Collaborative efforts in data collection and knowledge sharing between nations with shared profiles may foster more effective strategies.

  PublisherOA PDFDOI

• Understanding factors that impact vehicle travel times on urban road networks using machine learning approaches

  2025-01-01

  preprintOpen access

  While travelers may have the same origin and destination for their trips, these trips may be associated with different travel times. Understanding what factors contribute to different travel times contributes to better transportation planning and driver experience improvement. Big mobile device data offers new opportunities to understand travelers’ driving behaviors by providing high-resolution trip trajectory data. In this study we analyzed factors including driver behaviors, built environment and road characteristics, and external factors (e.g., traffic incidents and weather), for their contributions to travel times for two different travel contexts, commuting in an urban environment and rural trips involving a vacation destination, using machine learning approaches including time-series clustering and Random Forest regression models. We found that specific driver route choices (i.e., the different routes that drivers followed including the use of highways or local roads), and individualized travel speed behaviors including speeding contributed to variations in travel times. Vehicle profiles (i.e., types of vehicle drivers used) and road incidents had different impacts for different trip contexts. These results are useful for urban and transportation planners seeking mitigations that reduce travel time along heavily traveled driving routes and contribute to determining more accurate travel time predictions.

  PublisherOA PDFDOI

• Association Between Trust in Health Care Professionals and Health Care Access: Insights From an Online Survey Across 21 Countries

  International Journal of Public Health · 2025-04-10 · 3 citations

  articleOpen access

  Objectives: This study evaluates the association between trust in health care professionals and health care delays across 21 countries. Methods: We apply logistic regression models to survey data of over 621,000 individuals collected in Spring 2023. Results: Results show 44.5% of respondents with medical conditions experienced delays in accessing health care and 44.1% reported lack of trust in health care professionals. Those who trusted health care professionals had significantly lower odds of delaying medical care. Trust was most strongly associated with delays in the United Kingdom (OR = 0.373, 95% CI = 0.273-0.510), while South Africa had the smallest association (OR = 0.762, 95% CI = 0.582-0.997). Conclusion: Trust is important in influencing health care-seeking behaviors, though the causal direction warrants further research. There is a need for targeted strategies to build and sustain trust in health care relationships as well as enhancing health care access.

  PublisherDOI

• Examining trip-level errors in passively collected mobile device data for data quality assurance

  PLoS ONE · 2025-04-25 · 1 citations

  articleOpen access

  Location-based service (LBS) data passively collected by mobile devices has been widely adopted in multiple fields for its advantages in revealing travel behaviors. Data quality assessments have always been important steps for analyses using the data, but the impact of trip-level errors has not been a focus of these assessments. We examine a newly emerged type of error present at trip-level in LBS datasets that violates the spatio-temporal consistency of such data by including trips on road segments where and when there should be no trips. We designed a distributed-computing workflow to quantify the errors by comparing the number of trips on closed road segments during road closures with time periods before and after. Using two real-world cases from 2023, we examined multiple datasets acquired from major vendors in the US, and several of the datasets contained a significant number of trip-level errors. These findings point to the errors being present in recent datasets that have not otherwise been processed for data quality and can significantly impact analyses by data users. Data users should consider conducting trip-level error data quality checks as part of their preprocessing steps.

  PublisherDOI

• Assessing the legacy of redlining on spatial inequities in social and environmental determinants of health

  Applied Geography · 2025-04-18 · 4 citations

  articleOpen accessSenior author

  Although housing discrimination was outlawed in the United States in 1968, historic redlining remains a driver of racialized inequities in environmental health. However, there are many aspects of environmental health that are not yet well-understood in relation to redlining. We investigated the legacy of redlining on social and environmental determinants of health and the spatial distribution of these relationships across Baltimore and Philadelphia. We use publicly available spatial data sources on sociodemographics, built environment, housing, mobility, and arrests to understand the distribution of determinants of health given historic redlining. Multiscale geographically weighted regression was implemented to measure the relationship between these dimensions and redlining grades. While we identified strong, spatially heterogenous relationships between redlining and social and environmental determinants of health, for nearly all determinants of health, we observed the most adverse characteristics in “C” tracts, indicating a yellow-lining effect. Meanwhile, redlined tracts in both cities exhibited a mix of built environment characteristics, including higher levels of walkability, housing density, renter-occupied housing, and vacancies. Our findings suggest that while redlining has played a role in shaping neighborhood conditions, other factors, such as ongoing disinvestment and neighborhood transformation processes are likely influential in determining current social and environmental determinants of health. • Most adverse social and environmental determinants of health in yellow-lined (“C”) tracts. • GIScience methods are applied to investigate associations between redlining and urban built environments. • Strong, spatially heterogenous relationship between redlining and determinants of health. • Other racialized neighborhood change processes may attenuate the relationship between redlining and health.

  PublisherDOI

• Analyzing travel behavior differences across population groups: An explainable machine learning approach with big mobility data

  Journal of Transport Geography · 2025-07-25 · 3 citations

  articleSenior author
  PublisherDOI

• Mobility-GCN: a human mobility-based graph convolutional network for tracking and analyzing the spatial dynamics of the synthetic opioid crisis in the USA, 2013-2020

  arXiv (Cornell University) · 2024-09-16

  preprintOpen accessSenior author

  Synthetic opioids are the most common drugs involved in drug-involved overdose mortalities in the U.S. The Center for Disease Control and Prevention reported that in 2018, about 70% of all drug overdose deaths involved opioids and 67% of all opioid-involved deaths were accounted for by synthetic opioids. In this study, we investigated the spread of synthetic opioids between 2013 and 2020 in the U.S. We analyzed the relationship between the spatiotemporal pattern of synthetic opioid-involved deaths and another key opioid, heroin, and compared patterns of deaths involving these two types of drugs during this period. Spatial connections and human mobility between counties were incorporated into a graph convolutional neural network model to represent and analyze the spread of synthetic opioid-involved deaths in the context of previous heroin-involved death patterns.

  PublisherOA PDFDOI

• Does Context Influence How People Higher and Lower in the Propensity to Worry React to Uncertainty?

  2024-06-18

  preprintOpen access1st authorCorresponding

  It is well established that individuals high in chronic worry experience greater anxiety when faced with uncertainty than those lower in chronic worry (e.g., Dugas et al., 2004); however, there is a dearth of research on other potentially important cognitive and emotional reactions to uncertainty. The present dissertation examined whether individuals higher in worry have greater beliefs that uncertainty is unfair, they do not deserve to be uncertain, they are not able to cope with uncertainty, and anger, as well as whether individuals lower in worry have more positive appraisals (being confident they can handle uncertainty, accepting uncertainty, happiness, excitement, calm, feeling psychologically safe, and predicting uncertain scenarios will end well). This dissertation also examined whether context moderates reactions to uncertainty, as most clinical research has focused on decontextualized reactions. Study 1 examined the impact of being alone or with others in uncertainty, Study 2 examined whether uncertainty was expected or unexpected, and Study 3 examined whether the uncertainty was likely to end positively or negatively. It was predicted that individuals higher in worry would show less difference in reactions based on contextual factors, representing a general intolerance of uncertainty irrespective of context. Findings across the three studies suggested that chronic worry is associated with greater anxiety, greater perception of uncertainty, greater doubt about one's capacity to cope, low confidence handling uncertainty, and greater anger, rather than finding uncertainty unfair, feeling one doesn’t deserve to be uncertain, or not accepting uncertainty. Individuals higher in worry did not statistically experience less positive emotion in reaction to uncertain situations that are likely to end well, nor did they appraise uncertain positive scenarios as more likely to end negatively or as more unsafe than those lower in worry. On average, individuals higher in worry responded to context similarly to those lower in worry, finding being alone in uncertainty, not expecting uncertainty, and threatening outcome uncertainty to be more aversive. In summary, chronic worry appears to be associated with stronger negative cognitive and emotional reactions, especially anxiety and doubts about capacity to cope with uncertainty, to uncertain scenarios with a greater potential for threat.

  PublisherOA PDFDOI

Recent grants

• Epidemiology and transmission of subclinical malaria

  NIH · $8.6M · 2024

Frequent coauthors

• Lauren Berlant

  46 shared

• Junchuan Fan

  13 shared

• Kimberley W. J. van der Sloot

  Dialyse Centrum Groningen

  8 shared

• Hai Lan

  8 shared

• Frauke Kreuter

  6 shared

• Christina M. Astley

  Boston Children's Hospital

  6 shared

• Luliang Tang

  State Key Laboratory of Information Engineering in Surveying Mapping and Remote Sensing

  6 shared

• Tamara Cadet

  University of Pennsylvania

  5 shared

Education

• Ph.D., Geography

  University of California, Berkeley

  1992

• M.A., Geography

  University of California, Berkeley

  1988

• B.A., Geography

  University of California, Santa Barbara

  1985

Awards & honors

• UCGIS Fellow