Luyuan Shi

● 347-***-**** ● acy2cy@r.postjobfree.com ● www.linkedin.com/in/luyuan-shi

*** * ***** **, *** York, NY 10025

EDUCATION

Columbia University New York, NY

Master of Science, Statistics, GPA: 3.7 Sep 2015 - Dec 2016

• Relevant Courses: Data Mining, Big Data Analysis, Machine Learning, Advanced Data Analysis, Multivariate Statistical Inference, Stochastic Method in Finance, Linear Regression Models, Time-Series Modeling Shanghai Jiao Tong University (SJTU) Shanghai, China Bachelor of Science, Mathematics, GPA: 3.7 Sep 2011 - Jul 2015

• Honors: University Scholarships (10%), Outstanding Graduate (5%), Mathematical Contest in Modeling (20%)

• Relevant Courses: Mathematical Analysis, Advanced Algebra, Mathematical Statistics, Scientific Computation, Python, C Programming, Probability, Stochastic Process, Econometrics, Mathematical Finance PROFESSIONAL EXPERIENCE

Lighter Capital Seattle, WA

Intern, Data Scientist Jun 2016 - Sep 2016

• Pulled data from SQL, MongoDB and consolidated data from multiple sources. Performed intensive data cleaning and updated the mass of data into a digestible format that can be imported into statistical modeling with Python and Excel.

• Built ARIMA, SARIMA time series models and linear regression models to predict the revenue of 150 companies after intensive data cleaning and compiling. The model increased 12-month prediction accuracy from 70% to 98%.

• Developed classification methodologies using KNN, logistic regression as well as fundamental analysis methods with a focus on financial performance to filter potential clients and created a rating system of our clients.

• Designed an interactive visualization interface to generate financial data summary and model output with R Shiny and Tableau. China Huarong Asset Management Co., LTD Beijing, China Intern, Quantitative Analyst Jan 2015 - Aug 2015

• Responsible for collecting stock, bond price and financial data from WIND and targeted companies’ financial reports. Cooperated with marketing and financial team to keep the complex data consistency and integrity with Excel and MySQL.

• Performed valuation on companies using CAPM, multiple regression model to identify entry and exit points of equity positions.

• Conducted model validation and made adjustment to the coefficients of multi factor model according to the daily updated data. PROJECTS

Big Data Analysis of Million Song Nov 2016 - Dec 2016

• Generated data visualization with Tableau to find out the main trend of contemporary popular music on the Million Song Dataset.

• Transformed pitch and beats and other music attributes into mathematical features such as standard deviation, skewness and kurtosis. Applied Principle Component Analysis (PCA) to implement feature dimension reduction in Python.

• Set up Spark environment and ran K-means and Hierarchical Clustering methods with Spark MLlib package to cluster the music into four groups. Built word cloud on the clusters and characterized their genres as Pop, Jazz, Punk, and Trance.

• Built a web-app music recommendation system based on the highest similarity of audio attributes with Python and HTML. Analysis of Damage Awards in Jurisdiction Cases Oct 2016

• Utilized Decision Tree, Random Forest, Logistic Regression and Support Vector Machine(SVM) methods to build classification models on whether a jurisdiction case would receive damage awards.

• Performed Cross Validation to optimize the models and made comparison of different classification methods.

• Constructed Generalized Linear Model and Stepwise Regression models to select significant variables and make prediction on the amount of damage awards of plaintiff.

Data Science Project: US Census Marketing Campaign Dec 2015

• Applied generalized logistic regression model in 2010 US Census tract-level data to forecast Census mail return rate.

• Generated a heat-map visualization on US map to reflect the propensity score distribution of Census mail response across US.

• Advised an audience segmentation strategy for 2020 Census marketing campaign regarding to each tract’s socioeconomics, housing, marriage and education characteristics by K-medoids clustering approach. SKILLS & INTERESTS

• Software: Python (4 years), R/Shiny (2 years), SQL (2 years), Tableau (1.5 years), SAS, C++/C, Excel, Spark, Hadoop, Access

• Interests: Basketball, Photography, Swimming