Data Team Leader

Zeyu Dong

**-** **** **, ********, NY *****

646-***-****

******@********.***

Education

Columbia University, Graduate School of Arts and Science, New York, NY September 2014-February 2016 MA in Statistics, GPA: 3.8/4.0

Beijing Institute of Technology, College of Math, Beijing, China September 2010-July 2014 BS in Math, GPA: 3.3/4.0

Internship Experience

Findream, New York, USA December 2016 – Present

Data Analyst

Extracted personal financial information from different databases and conducted EDA and data mining methods to find potential credit pattern in user data

Identified the potential credit risks and fraud trends by establishing logistic regression model in Python with 75.4% accuracy

Improved LR model in Python through a linear combination of XGboosting, Random Forest (decision tree based), and SVM with 5.3% more accuracy

Visualized analysis results to senior management team to support data-driven business decision Arecy, New York, USA May 2016 – December 2016

Trainee

Established a real-time stock-analyzer platform with data ingestion layer (Kafka), data storage layer (Cassandra, NoSQL database), data computation layer (Apache Spark)

Created Zookeeper container in Docker machine to assist Kafka to fetch real-time stock price from Google Finance and Redis container to assist Kafka to filter data to dashboard

Designed front-end dashboard to visualize real-time stock analysis with node.js, D3.js, jQuery and JavaScript for clients Mass Mutual Financial Group, Hong Kong January 2013-June 2013 Management Trainee

Implemented table partitioning on SQL Server to make it more efficient and better resolution for data warehousing and reporting purposes.

Provided high quality financial data for back-end developers from different databases with SQL Academic Research

Sales forecast of Rossmann Company, Columbia University September 2015-December 2015 Team leader

Conducted exploratory data analysis (EDA) to summarize the characteristic of some features and perform feature engineering

Applied ARMA (time-series) model to predict the sales in R and enhanced the performance with Gradient boosting algorithm by 4.1%

Visualized both prediction results in R to support decision making process Sentiment analysis in movie reviews, Columbia University February 2015-May 2015 Team Leader

Transformed review data into numerical data in Python through TF-IDF/W2V and established several statistical models (Random Forest, XGboosting, CNN) to predict sentiment value

Improved the XGboosing algorithm performance by 5.8% with K-means clustering method in Python Mathematical Contest in Modeling, Beijing Institute of Technology February 2013-May 2013 Team Leader

Developed a web-scraping program in Python to automatically extract data from several websites and joined different tables by region

Conducted data imputation by KNN and established Gradient boosting model to predict load Skills

R (skilled), Python (skilled), SQL (skilled), javascript (intermediate), Hadoop MapReduce (basic), Apache Spark (intermediate), Pig (basic), Hive (basic)

Contact this candidate