Methods/Statistics
Towards a methodology for community-based spatial epidemiology: Lessons from the collection, geocoding, and analysis of unstructured location data within a community-based research setting Esteban J. Valencia* Esteban J. Valencia Kathleen Deering Shira Goldenberg Kate Shannon Hoaxuan Zhou Ofer Amram
Background
Spatial epidemiology within community-based research can uniquely advance knowledge on the social ecology of health for marginalized peoples, particularly within micro-level environments. Yet spatial methods require precise, structured location data. For community-based researchers using self-report tools that collect unstructured location data, there is no standard approach for data collection and geocoding. To address this, we review the collection, geocoding, and analysis of unstructured location data from community-based research with marginalized women. We further synthesize these processes into a framework for collecting and geocoding unstructured location data.
Methods
We draw on two community-based cohort studies of a) sex-workers and b) women living with HIV, from Vancouver, Canada (2010-2023). Location data were elicited using self-report, and include: place of residence, overdose events, healthcare access sites (e.g., HIV care), and experiences of violence. We detail an approach to parse and geocode unstructured location data and evaluate the quality of resultant coordinates. We used these data to create variables representing novel integrations of socio-structural and spatial components.
Results
Median participant age was 35 years, 38% were Indigenous, and 66% were women of color. Over 13 years, 981 participants contributed 89207 location data points, 81% of which were successfully geocoded. Almost half of all coordinates corresponded to one of 300 distinct sites in Vancouver. Data were used to create socio-structural variables, such as residential proximity to spatiotemporal clusters of non-fatal overdose or experiences of gender-based violence.
Discussion
For community-based researchers who rely on self-report tools, we provide a unique methodological framework demonstrating how the collection of unstructured location data can be used to understand intersections between spatial epidemiology and the social ecology of health for marginalized peoples.