Baiyan Ren
adiw0w@r.postjobfree.com 917-***-**** LinkedIn: www.linkedin.com/in/baiyanren
Github: https://github.com/BaiyanRen Kaggle: https://www.kaggle.com/baiyanren Summary
{ Three years of hands-on experience in dealing with both structured and unstructured data, performing data wrangling and visualization, and applying machine learning approaches to business problems
{ Experienced Ph.D. researcher solving complicated problems, designing research process and managing projects Skills
{ Data Science:
- Programming: R, Python, SQL
- Data analysis: Microsoft Office, Tableau (Desktop Specialist Certificate), Power BI
- Machine learning: Regressions, Decision Trees, Support Vector Machines (SVMs), ensemble methods
{ Research: Experimental design, project management, scientific writing and public speaking. Education
{ University of Nebraska Medical Center, Omaha, NE Expected Aug. 2021 Ph.D. in Neuroscience GPA: 3.5/4.0
{ Udacity
The School of Data Science, Data Analyst Nanodegree Nov. 2020
{ China Pharmaceutical University, Jiangsu, China Jul. 2016 B.S. in Pharmaceutical Science GPA: 3.3/4.0
Relevant courses: Linear algebra, Probability, Biostatistics, Data structure and algorithm, Machine Learning Projects
{ Predicting Medical Appointment Attendance
- Applied supervised machine learning to predict appointment absence
- Performed univariate, bivariate, and multivariate explorations on 110,527 medical appointment records containing 14 features (age, gender, financial and physical conditions, etc.) of patients; conducted features selection, algorithms selection, and parameters tuning using Python statsmodel and scikit learn libraries
- Built Random Forest Classification to predict appointment absence using age, physical conditions, and financial support, with accuracy 0.66, precision 0.31, recall 0.55
- https://github.com/BaiyanRen/Medical-appointment-absence-prediction
{ A/B Testing on Webpage Design
- Analyzed 290,584 A/B test records collected in 3 weeks run by an e-commerce website
- Built Logistic regression model to assess whether page view, living country, and their interactions could predict whether a user decides to pay for the product (convert); performed hypothesis testing to analyze whether the new page increases the conversion rate
- Reported that page view and country do not predict the conversion of users, avoiding unnecessary expense on launching the new webpage
- https://github.com/BaiyanRen/AB-test-of-e-commerce-website
{ Modified SEIR Model for COVID-19 Cases Prediction
- Built a modified SEIR (Susceptible - Exposed - Infectious - Recovered) model using Python to forecast the COVID-19 cases
- Processed daily records of COVID-19 in the US, France, and Italy from global data; modified the SEIR model considering infectious viral carriers occasionally do not exhibit symptoms; applied Least Square to optimize the parameters: infection rate, recovery rate, exposed rate, and reinfection rate, in the modified SEIR model
- This model fits the real-world COVID-19 data in the US, France, and Italy very well
- https://www.kaggle.com/baiyanren/modified-seir-model-for-covid-19-prediction-in-us Experience
Research Assistant, University of Nebraska Medical Center, Omaha, NE Aug 2016 - now
{ Determined cell cycle dynamics of human stem cell-derived astrocytes programmatically; Analyzed and visualized proteomics profiles of human stem cell-derived astrocytes using Python
{ Delivered presentation on global and regional conference: Neuroscience 2018 and Neuroscience 2019, Midwest Student Biomedical Research Forum in 2017, 2018, 2019, and 2020