Amber (Biying) Tan
I'm a highly motivated and passionate Data Scientist whose dream is to use my talents in machine learning/AI and statistical analysis to solve real-world problems and make the customer achieve more.
Vancouver, Canada
Work Experience
Data Scientist II, Electronic Arts (EA), Jul 2017 - Present
- Apply quantitative analysis and machine learning to detect in-game exploits, fraud and other security threats in Hockey Ultimate Team (NHL HUT) with Python
- Built hybrid classification models with Random Forest and Isolation Forest to detect fraudulent accounts and anomalous behaviors including coin farming/selling, credit card fraud, match cheating to help reduce fraud losses by millions of dollars annually
- Build Dashboards to monitor the health of the game economy, analyze data anomalies and communicate findings to the team with Tableau and SAS
Data Scientist, PHEMI Systems, Aug 2016 - Jun 2017
- Built PHEMI Data Science Toolkit to perform descriptive analysis and visualizations to showcase data insights with Spark and D3.js
- Developed text mining pipelines in Spark to analyze large unstructured datasets with Latent Dirichlet Allocation (LDA), text classification and sentiment analysis
- Created training materials for the client about Machine Learning models, Natural Language Processing and Recommender System
Data Analyst, Move Inc, Jan 2016 - Aug 2016
- Improved the accessibility of multi-sourced data across the team by designing and building a high-performance data warehouse on AWS to ingest data from various data sources
- Reduced the data and content development time by more than 50% by developing an ETL pipeline to streamline the data transformation, aggregation and reporting process
- Built Dashboards to extract and visualize key trends impacting residential real estate using Tableau
Data Analyst, SingTel, Jun 2014 - Nov 2015
- Improved user engagement for Newsloop (Top 5 News App in Singapore) by 89% by developing a real-time personalized push notification system using Latent Dirichlet Allocation (LDA) and topic modeling
- Optimized the game acquisition strategy by building a classification model with Logistic Regression to predict viral games with 85% accuracy
- Supported Games and News Divisions’ data needs by developing ETL and reporting pipelines
- Defined metrics and built dashboards to gain insights into product and user behaviors using Tableau
Education
Master of science, Data Management & Analytics, Information Systems
Singapore Management University, 2012 - 2014
Specialized in Social Network Mining and Natual Langauge Processing
Bachelor of Engineering, Computer Science
Sichuan University, 2008 - 2012
Academic Scholarship in the consecutive 3 years for Top 5% students for academic achievement
Skills
- Machine Learning: Classification, Clustering, Regression, Fraud Detection, Natural Language Processing, Recommender System, Social Network Mining
- Statistics: Probability, Distribution, Hypothesis Testing, A/B Testing, Regressions, Time Series Analysis
- Programming languages: Python (SciKit-Learn, Pandas, NumPy), Scala, R, Java
- Database: MySQL, Amazon Redshift, SQL Server
- Big Data: Spark (MLlib), Hive, Hadoop, MapReduce
- Visualization: Tableau, Python (Matplotlib, Seaborn), SAS, D3.js
Publication
- Biying Tan, et al. "Clairvoyant-Push: A Real-Time News Personalized Push Notifier using Topic Modeling And Social Scoring for Enhanced Reader Engagement." IEEE International Conference on Big Data, 2015
- Biying Tan, et al. "Online Community Transition Detection." Web-Age Information Management (WAIM2014). Springer International Publishing, 2014. 633-644
- Du Juan, Biying Tan, et al. "Social Listening for Customer Acquisition." Social Informatics (SocInfo2013). Springer International Publishing, 2013. 75-80
- Biying Tan, et al. "Detection of high-risk zones and potentially infected neighbors from infectious disease monitoring data." Database Systems for Advanced Applications (DASFAA2012). Springer, 2012
