Shiyuan Chen
**** ***** ****** **, ************, PA, 19104
215-***-**** adc9nz@r.postjobfree.com www.linkedin.com/in/shiyuanchen9601 EDUCATION
Drexel University LeBow College of Business GPA:3.8/4.0 Philadelphia, PA Master of Science in Business Analytics March 2020 Relevant Courses: Statistics for Business Analytics, Data Mining (R), Database Analysis and Design Business (Tableau & SAP Hana), Multivariate Analysis, Risk Management
East China University of Science and Technology Shanghai, China Bachelor of Science in Business Administration June 2018 Relevant Courses: Multivariate Statistics (SPSS), Probability Theory, Mathematical Statistics TECHNICAL SKILLS
Tools and Programming: R, SQL, SAS, Power BI, Tableau, SAP Hana, Advanced Microsoft Excel, Python, SPSS Digital Analytics: Google Analytics, Google Adwards, Digital Marketing, Market Research PROFESSIONAL EXPERIENCE
Peco Energy Company Philadelphia, PA
Practicum Data Analyst September - December 2019
• Handled large dataset with 10K+ observations and 15 variables with Python.
• Used ARIMA and Holt-Winter to develop time series models for different types of business that forecast call volumes for the call center which achieved 75% accuracy.
• Predicted shrinkage and abandon rates with linear regression and achieved 85% accuracy on testing set.
• Improved the company’s methodology by introducing new features and increasing granularities to more accurately predict call center labor needs.
Donghai Securities Qingdao, China
Business Analyst Intern June - August 2017
• Utilized internal CRM system to segment customers, used MySQL to identify the top client group and provided different targeting strategies and product recommendations.
• Performed a text massage campaign to introduce wealth management products and service to customers which improved conversion rate by 30%.
• Analyzed transactional database to provide insights to the stock trading and securities investment departments. Industrial and Commercial Bank of China (ICBC), Chengyang Branch Qingdao, China Data Analyst Intern January - March 2017
• Built linear regression model using defect rate, late payment rate and other basic informations to serve as features of models like fraud and credit score.
• Designed A/B testing for marketing campaigns and improved response rate by 10%.
• Analyzed customer behavior and partnered with third party on credit card benefits for targeted consumer group.
• Organized customer demographic and financial information into a database. ACADEMIC PROJECTS
Identifying Potential Subscribers for Long-Term Deposit (R Studio) (2019)
• Used SVM to identify potential subscribes for long-term deposit in a dataset with 45.2K observations and 17 variables.
• Performed 10-fold cross-validation to test the model performance; achieved 83% accuracy rate and 85% recall rate.
• Provided different feasible strategies for banks by controlling classification weights in SVM. Portfolio Optimization Based on Modified Markowitz Method (SAS) (2019)
• Improved the traditional Markowitz formula, introduce Buy-in threshold constraints, Round Lot Perchasing constraints and Diversification constraints to make it suitable for modern investment.
• Used Mixed Integer Nonlinear Programming in SAS to build the model to calculate the optimal portfolio of eight stocks and a risk-free bond; the monthly return of the portfolio reached 2.3%, and the risk variance was controlled at 0.22%. Screening Tool for Chronic Kidney Disease (CKD) (R Studio) (2019)
• Processed 6000+ patients information dataset and used K-NN and linear regression to fill in the missing data.
• Used clustering analysis and factor analysis to perform feature selection, identifying 12 predictors among 39.
• Built logistic regression model to predict potential kidney disease patients in R, and provided medical workers with a simple screening tool with an accuracy rate of 86%. Sentiment Analysis of Muisc App Reviews (R Studio) (2018)
• Cleaned, categorized and performed data mining on 90,000+ music reviews; built new sentiment scoring form.
• Used R to create a new lexicon and achieved 85% accuracy when testing customer sentiments.