Chang Xu (Grace)

347-***-**** adg65h@r.postjobfree.com 504 W 110th St, Apt 8A, New York, NY 10025

PROFESSIONAL SUMMARY

Graduate with background in mathematics and statistics, and expertise in various statistical software and programing languages. Seeking to leverage strong analytical skills in an intellectually stimulating role. EDUCATION

Columbia University: Graduate School of Arts and Sciences New York, NY Master of Arts - Quantitative Methods in the Social Sciences GPA 3.5 GRE Quant: 170/170 May 2020 Sichuan University: Department of Mathematics Chengdu, China Bachelor of Science - Mathematics (concentration: Statistics) June 2018 RELEVANT COURSEWORK AND SKILLS

Data Science: Advanced Machine Learning, Data Mining, Natural Language Processing, Social Network Analysis, Data Visualization Statistics: Bayesian Statistics, Time Series Analysis, Experimental Design, Sample Surveys, Applied Regression, Multivariate Statistics Programming/Computer: R, Python, STAN, SQL, RShiny, MATLAB, SAS, LaTeX, markdown, Microsoft Office, Adobe (Ps, Pr) RESEARCH EXPERIENCE

Research Assistant Institute for Social and Economic Research and Policy, Columbia University Sept 2019 - May 2020 o Constructed multiple ETL pipelines to manage historical NYC micro census data and GIS information in real time o Developed and optimized string cleaning functions to maintain tidy address data o Conducted data transformation with R on unstructured name and address datasets to build historical NYC street name dictionaries o Set rules and filters to detect and report error and incomplete data entries o Utilized GitHub repository for version control and team collaboration Research Assistant Sociology Department, Columbia University Sept 2018 - Dec 2018 o Conducted literature reviews on indie literary presses to explore press trajectories, editor-writer relationship, and distribution models o Standardized sampling criteria and performed stratified sampling in R to generate a selection of presses and books o Extracted relevant sample data as nodes and ties for social network analysis SELECTED ACADEMIC PROJECTS

Master's Thesis - Motherhood Penalty and Gender Pay Gap Mar 2020 - May 2020 o Designed research project to evaluate the gender pay gap and assess trends over time o Selected variables and extracted data from General Social Survey and U.S. Census Bureau for data visualization o Built regression models in R to examine factors that influence personal income, emphasizing marital and parenthood status Image Style Transfer Group Project Nov 2019 - Dec 2019 o Implemented neural style transfer algorithms using TensorFlow to mix content and style images in Python o Experimented with layer selection, weight assignment, and train loss variation for optimal results o Wrote report on research design, methods, and results, and conducted in-class presentation Facial Expression Detection Group Project Oct 2019 o Constructed subtle facial expression detection and classification program in R to distinguish 20 refined distinct facial features o Implemented and tuned various supervised learning models in R (LDA, random forest, GBM, etc.) o Identified most accurate models via confusion matrix, with test accuracy of 58% NYC Shooting Crime Map Group Project Sept 2019

o Integrated NYC shooting incidents into Shiny app, and added filters such as of time, location, and victim backgrounds (Link) o Created interactive map in Leaflet, indicating crime density by color and clustering circles o Analyzed crime data and visualized results in various plots (ggplot2) Lyrics Analysis Apr 2019

o Analyzed lyrics of 1 million songs with sentiment analysis, topic modeling, and TF-IDF in Python, to evaluate changes over time o Created a lyric generator with Markov Chain Monte Carlo methods studying 75 Bowie songs, reaching a similarity of 40% Bachelor's Thesis - Tax Revenue Modeling and Forecasting Apr 2018 - June 2018 o Built ARIMA and GARCH time series models to detect irregularities in daily tax revenue, and make risk predictions o Examined ACF, PACF, and implemented Tukey's HSD to check stationarity