In this section of my site, I have included samples of my work in the form of course final projects. As an area of interest of my advisor, many of my final projects include data related to the United States opioid epidemic.
For this course, I utilized PostgreSQL (still commonly referred to by the name of its predecessor "POSTGRES") for database development to combine data from 2 distinct sources: the Opioid Environmental Policy Scan and County Health Rankings & Roadmaps datasets. I utilized the PostGIS spatial database extender to uniquely preserve geospatial information in the data. Raw data were extracted, transformed, and loaded (ETL) via a combination of R, python, and SQL scripts. Following data storage, I utilized Tableau to create an interactive data dashboard to visualize patterns in this data.
For this course, I utilized a Virginia county-level subset of data from the database created for my Scientific Databases course (described above) along with opioid overdose related death counts obtained via the Centers for Disease Control and Prevention (CDC) Wide-ranging ONline Data for Epidemiologic Research (WONDER) tool. Data preprocessing was completed in R and linear mixed-effects models, via the Linear and Nonlinear Mixed Effects Models (nlme) R package , were used to identify cross-sectional and longitudinal measures associated with longitudinal opioid overdose related death counts.
For this course, I utilized Virginia state-level economic measures of monthly state unemployed persons counts and monthly noninstitutionalized population (as a covariate) obtained via the United States Bureau of Labor Statistics (BLS) , to forecast monthly Virginia opioid overdose death counts. Data preprocessing was completed in R and dynamic harmonic regression models, via the forecast R package , trained on data from January 2000 to December 2019 were used to forecast opioid overdose related death counts for the entire year of 2020 (January 2020 to December 2020).