Data/Risk Analyst

LUQI TANG

617-***-**** ******.****@*******.*** Boston, MA

Summary

Data Analytics candidate with 5-year academic experience in statistics, model building, and data visualization.

Technical Skills: 4-year SAS experience, 3-year R experience, proficient in SQL, Python, SPSS, Tableau, Power BI, Google Analytics, Microsoft Office (Access, Excel, PPT, Word)

Statistical Methods: Linear Regression, Logistic Regression, Poisson Regression, Time Series Regression, Decision tree Education

Boston University, Boston, MA 01/2020

Master of Arts in Statistics GPA: 3.43 / 4.0

Courses: Stochastic Process, Linear Models, Generalized Linear Models, Data Science in R, Database Design and Implementation for Business, Mathematics of Financial Derivatives Tamkang University (TKU), New Taipei City, Taiwan 06/2018 Major: Bachelor of Science in Statistics Minor: Banking and Finance GPA: 94.59 / 100

Courses: Probability, Mathematical Statistics, Survey Sampling, Categorical Data Analysis, Design of Experiments, Nonparametric Statistics, Time Series, Money & Banking, Financial Management, Econometrics, Financial Statements Analysis, Advanced Analysis of Financial Issues

Scholarship for Academic Excellence, top 1 out of 180 (7 times); Honors Program (Top 5% of Statistics Department)

3rd Prize in the Tenth Questionnaire Design and Statistical Data Analysis Competition Experience

Data Analyst Intern (SAS), Joblogic-X Corporation, Houston, TX 11/2019-02/2020

Aggregated and cleaned data from TransUnion on thousands of customers' credit attributes; Performed missing value imputation using population median, check population distribution for numerical and categorical variables to screen outliers and ensure data quality

Leveraged binning algorithm to calculate the information value of each attribute to evaluate the separation strength for the target variable; Built logistic regression model to predict the probability of default; Used stepwise selection method to select model candidate variables

Checked variable multicollinearity by calculating VIF across predictors; Tested multiple models by switching variables and selected the best model using performance metrics including ROC, KS and Somer’s D Investigations On Aviation Accidents (R), Boston University, MA 01/2019-05/2019

Collected data from NTSB and randomized sample into 46,581 observations; Narrowed down the scope to 3 aspects that would affect air crashes: Aircraft, Weather and Crew; Used EDA and Stepwise Algorithm to select variables

Created a parametric model which is a combination of a logistic and Poisson regression model based on the range the response variable falls in; Performed Cross-Validation for non-parametric classification tree

Showed some combinations of the number of engines and aircraft damage levels are riskier than others, which airlines should pay more attention to decrease aviation accidents Indian Liver Patient (SAS, R), Tamkang University, Taiwan 09/2017-12/2017

Converted variables and handled missing data; Built logistic regression model, used Hosmer and Lemeshow goodness of fit to detect whether the model fits observations or not and used ROC curve to see model's discriminatory power; Undertook model selection and validation

Created a logistic regression model to predict the probability of people who will suffer from liver diseases Conversion between TWD and CNY Cash Rate-Selling Rate (R), Tamkang University, Taiwan 02/2017-06/2017

Organized members to establish the best SARIMA model based on the CNY cash selling rate from 01/01/2015 to 03/31/2017, totaling 821 records (including 10 predicated records) and validated predictor values to the actual ones

Considered the domestic and global economic environment, drawing the conclusion that the exchange rate of RMB against NTD should gradually increase and it is suitable to purchase RMB systematically and periodically The rate of violent crime in the United States (SAS, R), UCI & Tamkang University, Taiwan 09/2016-12/2016

Led a group of 9 in researching elements leading to violent crime in the USA; Assigned a reasonable division of labor and set deadlines for other team members; Submitted a 50-page analytical report as well as presented to 70 people

Established a multiple linear regression model based on 8 variables offered by UCI open database; Undertook 4 model selection algorithm (all enter, forward selection, backward elimination, stepwise regression), validation and diagnostics(residual analysis, diagnostics for heteroskedasticity, autocorrelation test, normality test)

Predicted crime rate based on race, age, income, educational background, marriage, family background, structure and density of population

Contact this candidate