The interdisciplinary work at the core of this research relies on the exploration of larger parameter spaces to span traditional distinctions between qualitative and quantitative research. Whereas earlier research in humanities and social science computing concentrated on including more cases, much of the new paradigm in computation relies on operations on the parameter space to create significant findings for the outlying factors within a complicated model. This allows a higher level of abstraction that was once reserved for qualitative research, but also demands a new approach to data collection, coding, and visualization.
In collaboration with other investigators and community partners, Dr. Price’s research seeks to understand the underlying causes of health disparities across different social subgroups by employing machine learning techniques to build cross-functional relationships between socioeconomic, atmospheric, geospatial, and biomedical data. The combination of challenging compute processes and large datasets requires the use of high performance computing. The collaborative and interdisciplinary process requires iterative visualizations of how changes in the operational space transform representations from diverse functional spaces so that a unified model can be developed and tested. For example, if the modeling for air pollution operates on a different time scale from the models for health effects, which are in turn different from those for health disparities, then a combined space requires larger matrix operations, more parallelization, and more efficient coding. Collaborators need quick turnaround to see how small model changes effect the shared conceptual space, and to further refine the modeling assumptions and design appropriate tests.