Data Analyst, python, SQL, machine learning, AB test

Kuo Song

Jersey City,***** +1-347-***-**** *********@*****.***

TECHNICAL SKILLS

• Programming: Python (Scikit-learn, TensorFlow, NumPy, Pandas), R, MySQL, Tableau, Hadoop (Hive), Spark, SPSS, C++

• Machine Learning: Logistic Regression, Random Forest, Boosting Tree, SVM, Clustering, KNN, Recommender System

• Analysis Techniques: Hypothesis Testing, Exploratory Data Analysis, Time Series Analysis, Web Crawling, AWS, GCP EDUCATION

Fordham UNIVERSITY New York, USA

Master of Science in Business Analytics (Data Science), GPA: 3.9/4.0 Aug. 2019-Dec. 2020 Courses: Data Structure, Time Series, Web Analytics, Deep Learning, Machine Learning, Database Management Beijing Forestry University Beijing, China

Bachelor of Science in Finance and Banking (GPA:3.7/4.0) Sep. 2014-June 2018 PROJECTS

Using Image Classification to Enhance Yelp’s Rating System

• Tested Yelp rating system bias on restaurants with not enough reviews, and designed an algorithm that mines embedded information in images to boost Yelp rating performance

• Quantified images information by predicting the probability of being good quality for each image by a fine-tuned LeNet CNN model to acquire key variables that improve the predictive power of the original rating system

• Fed images and the 34 other preprocessed variables into Random Forest models, increasing rating accuracy by 12.8% Retrieval Based Twitter Customer Service Chatbot

• Build a retrieval-based customer service chatbot intended to provide the best possible service to customers

• Encoded questions with TFIDF, LSI, LDA, and Word2vec to calculated sentences’ matching scores to identify the most appropriate answer to the input question

• Designed user interface and set threshold similarity to filter questions and call manual intervention; tested and selected the best algorithm based on the input length, reducing human efforts by 58% compared with the original situation Empirical Study of Corporate Green Bond and its ESG Performance

• Explored the relationship among the issue of corporate green bonds, the ESG performance and firm valuation

• Collected and preprocessed data from Bloomberg, fed winzorized data into regression model and LDA model PROFESSIONAL & ACADEMIC EXPERIENCE

Fordham University, Science Center for Digital Transformation New York, USA Data Scientist June 2020-Sep. 2020

• Designed an environment-based agent, and built a machine learning pipeline with multi-layer perception and deep Q- learning model to support decision-making on stock trading by historical market data

• Responsible for collecting and preprocessing 10-year stock data, and performed deep exploratory data analysis to adjust trade strategy based on the Moving Average Convergence Divergence (MACD)

• Detected strategy decisions and the evolution of the total reward over episodes, reaching the model accuracy of 82% Kaiyuan Securities Co., Ltd Beijing, China

Data Analyst Dec. 2018-Feb. 2019

• Participated in a securities issue project of 5 million dollar, preprocessed structured and unstructured data of the issuer, and built random forest and linear regression models to evaluate enterprise values and business risks

• Updated securities information weekly based on the market indicators, performed data engineering to ingest millions of data records including data cleaning, creating time-series & dummy variables

• Monitored and interpreted anomalous market changes, and derived business insights through visualizations Chinalin Securities Co., Ltd Beijing, China

Data Analyst July 2018-Sep. 2018

• Conducted metric analysis over model result to analyze relationship between market return and seasonal impact

• Responsible for data management, analysis, and presentation, using MySQL, Python, and Tableau

Contact this candidate