Ruoxi Zhang

New York, NY, ***** 631-***-**** adbh0j@r.postjobfree.com https://www.linkedin.com/in/ruoxizhang218 EDUCATION

Stony Brook University Aug 2016 – Dec 2019

-Master of Science, Statistics Stony Brook, NY

• Cumulative GPA: 3.53/4.0

• Relevant Coursework: Data Analysis, Failure and Survival Data Analysis, Categorical Data Analysis, Machine Learning, Statistical Computing, Theory of Database, Stochastic Models, Probability Theory, Design and Experiment

-Master of Art, Public Policy

• Cumulative GPA: 3.58/4.0

Zhengzhou University Aug 2012 - Jun 2016

Bachelor of Art, Public Administration (Minor in Finance) Zhengzhou, China

• Cumulative GPA: 3.4/4.0

SKILLS

Coding: SQL, R (dplyr, caret, ggplot2), Python (Pandas, Numpy, Seaborn, scikit-learn), MATLAB, SAS, Stata Analytics Tools: AWS Redshift, Tableau, Excel

Machine Learning Technique: Regression, Classification, Clustering, Feature Engineering, Data manipulation and visualization EXPERIENCE

Healthfirst – New York, NY Intern, Data Science Jun 2019 – Dec 2019

• Member churn prediction: Provide risk scores for Medicare enrollee’s likelihood to churn which has a financial and Medicare Star impacts. Collect data from AWS Redshift using SQL, applied feature engineering and ML models including Logistic Regression, LightGBM and Random Forest using Python. Achieved 0.62 AUC and 0.79 Recall on the test dataset.

• Mental health progression prediction: Predict hospital admission risk (moderate risk) and suicide ideation risk (sever risk) caused by depression among members from 2016 to 2018. Identify members who have high risk to allow Healthfirst to prioritize follow-up and referral for intervention. Collect data from AWS Redshift using SQL, applied feature engineering and XGboost and Random Forest models using R, achieved 0.89 AUC, 0.88 Recall on the test dataset.

• Opioid misuse or overdose detection: Using R to calculate opioid levels and identify patients at risk of opioid misuse or overdose. Extract prescription drug claims data from AWS Redshift using SQL, calculate patients’ average daily morphine equivalent dose (MED), identify risk group and who appear to be doctor shopping based on Toolkit from HHS. Visualize outcome with Tableau dashboards.

Stony Brook University – Stony Brook, NY Assistant Research Analyst Jan 2019 - May 2019

• Establish and improve statistical methods in Teacher Retention Analysis including LASSO, K-fold cross-validation, Logistic Regression and Mediation Analysis using R.

• Assist visualization work including tables and charts using Latex in Teacher Retention Analysis. Bridge Trust – Zhengzhou, China Quantitative Analyst Intern May 2018 - Jun 2018

• Using Information Ratio as a criterion, given Price Earnings ratio on the individual company and the ratio in its corresponding industry. Build a model using MATLAB to identify companies whose stock price could increase in the future from a total of 895 companies.

• Build a model by MATLAB to simulate the industrial stock price trend given average quantiles of Price Earnings Ratio on individual companies

• Help on data visualization using Excel in the Annual Report of 68 trust companies in China. PROJECTS

Predict Wine Goodness from Review: Implement Ridge Regression using Python. Predict the points the wine should get based on its review text. Achieved 1.12 RMSE with Leave-One-Out-Cross-Validation. Rank 25th among 73 teams. Identifying Fraudulent Activities: Implement feature engineering on transaction data and build Random Forest model using Python to predicts whether a user has a high probability of using the site to perform fraudulent or not. AUC: 0.83. Precision: 0.96. Recall: 0.96

Larynx Cancer Survival Analysis: Build proportional hazards models using SAS on larynx cancer data and estimate hazard ratios with different disease stages. Using Chi-square goodness of fit test to find the best suitable model.

Contact this candidate