Marne Silk

Master's

Economics

San Francisco State University College of Business

Applied Research

• Student debt and health policy research: Built econometric and machine learning models to understand the impact of student debt on health outcomes (predicted probability theory, decision trees and linear discriminant analysis). Performed literature reviews and written analysis to support research hypotheses • Comparative analysis of environmental degradation: a collaborative project that predicted climate degradation as a function of macroeconomic variables using advanced time-series analysis (multivariate linear estimation, vector auto-regression, vector-error correction model, testing for co-integration and Granger causality) • Predicted employee absenteeism in transportation and logistics using machine learning techniques-logistic regression, random forests and decision-based methods, which aimed to determine if certain explanatory variables have a statistically significant relation to employee absent time as measured in hours • Survey and response analysis: a 5-month research project in collaboration with a SF Bay Area Chinese health coalition partner on the motivating factors of problem gambling among bus patrons within socio-demographic populations. Built a customized survey analysis; used Stata to execute statistical analysis and testing in order to analyze the effects and motivating factors of problem gambling. Made several actionable policy recommendations. • 2+ years of experience solving analytical problems and building models using quantitative, statistical and machine learning approaches • Experience performing data extraction, cleaning, manipulation and analysis using medium to large data sets • Advanced proficiency with statistical methods such forecasting, time-series, hypothesis testing, classification and multi-regression analysis • Advanced programming in R; experience writing SQL queries • Experience with data visualization libraries such as ggplot2, fpp2 and scientific computing and analysis packages such as dplyr and caret