• Used statistical models such as regressions and survival models to examine the impact of various factors on employee attrition. Generated attrition predictions using machine learning models.• Leveraged NLP and LLM techniques for sentiment analysis of employee surveys, summarizing survey comments into key topics for executive review and action. The results were reported to the leaders for strategic actions.• Automated data pipeline for weekly reports using Python. Improved work efficiency by 50 percent.• Cleaned external data, such as inflation and unemployment data, for use in statistical models.

• Conducted Pay equity analysis and identified the best remediation strategies to improve pay equity. Cleaned data, ran analysis using regression models, and generate remediation strategies simulation using R. Local teams used the remediation strategies to mitigate pay equity risk. • Lead a data science project by creating R Shiny apps that can predict salaries. Developed ETL process by extracting internal data using SQL and writing production codes to feed data into supervised learning models using R and Python. The apps are widely used by roughly 200 Talent Acquisition and Talent Management team members to estimate the salaries for new hires and promotions.• Created data visualizations by building dashboards. Extracted internal HR data and used Power BI and Tableau to show various HR metrics. The dashboards are used by HR leadership to devise strategies to improve HR policies.• Conducted several quasi-experimental analyses such as examining the effectiveness of handheld devices on productivity and retention and examining the effectiveness of a health program on revenue and turnover rates. Cleaned data using R and built causal inference models such as difference in differences, propensity score matching, and synthetic control method. The results of the projects are presented to senior leaders so as to offer guidance to critical business decisions.• Assisted other functions such as marketing teams or HR teams by helping them to implement causal inference models such as synthetic control method and instrumental variables method.• Extracted over 10 GB of data from various external data sources to improve models. Wrote codes in R and built data input from various sources such as Google Maps and labor market data using API. • Improved the efficiency of the apps and data science projects. Benchmarked different approaches and used parallel programming to optimize the coding speed. Cut down the running time of the apps by 50 percent.

• Evaluated the impact of TV advertising on the quantity demanded of e-cigarettes using AC Nielsen data. Used longitudinal data analysis and survival analysis in the research.• Examined the price elasticity of e-cigarettes of different brands using scanner data. Used causal models such as instrumental variable regression in the research.

• Constructed advanced econometrics models such as survival analysis, panel data model, logistic regression, instrumental variable regression to answer causal inference questions related to labor and health economics • Quantified the heterogeneous effects of maternal employment on adolescents’ smoking and drinking behaviors• Used public datasets such as the American Time Use Survey (ATUS) and the National Longitudinal Study of Adolescent to Adult Health (Add Health) to answer research questions• Created summary statistics and data visualization for research projects• Managed and organized over a million observations in ATUS using R and over 34,000 observations using Stata in Add Health• Presented and communicated research results in the Southern Economic Conference and Georgia State University Health Brown Bag Seminar

• Managed and organized datasets using Stata; wrote a programming package for several research projects in Education Economics • Provided summary statistics and data visualization for research projects• Communicated with Georgia Department of Education and collected data from different sources such as Georgia Department of Education and Georgia County Guide

• Searched various sources of data that contain the corresponding industry sectors for the companies• Scrapped data using R from various APIs such as Freebase API, Feedzilla API, Factual API and Wikipedia API• Experimented various statistical tools such as LASSO to increase the prediction accuracy of the application • Wrote documentations for the usage of the application• Prepared PowerPoint slides to report the progress of a newly developed product to senior managers and other team members