JMIR Publications

ABSTRACT

Contextual variables representing the economic, political, or cultural characteristics of a specific area have crucial applications in public health research and program evaluation, such as evaluating policy implementation in local areas or explaining variability in health outcomes across populations. However, accessing context-level data can pose significant challenges in the absence of monitoring systems. Even though the Internet can serve as a major source of information, website data is often unstructured and not suitable for analysis. This study aims to describe a novel research method that integrates web scraping, text mining, and spatial overlay analysis to convert unstructured website data into theoretically informed contextual variables. The paper is structured as follows. In the first section, we describe the method while introducing the techniques of web scraping, text mining, and spatial overlay analysis. The process is explained step-by-step and applied to a real research case to generate contextual-level variables on health assets with the potential to foster social connections among older adults in the context of a large regional public health program. The method, however, can also be useful in public health, health services research, health policy analysis, program evaluation, epidemiology, and other disciplines with an interest in contextual-level data where data is scarce, hard to obtain or reflects emerging issues where data has not been generated.