Candidates About

 

Caroline Xue 

SUMMARY:

·         Over 6 years of profound experience as a Data Scientist with excellent Statistical Analysis, Data Mining and Machine Learning Skills.

·         Worked in the domains of Financial Service, Healthcare and Retail.

·         Expertise in managing full life cycle of Data Science project includes transforming business requirements into Data Collection, Data Cleaning, Data Preparation, Data Validation, Data Mining, and Data Visualization from structured and unstructured Data Sources.

·         Hands on experience in writing queries in SQL and R to extract, transform and load (ETL) data from large datasets using Data Staging. 

·         Proven ability in using Text Analytics and statistical modeling techniques such as: linear regression, LASSO regression, logistic regression, elastic net, ANOVA, Monte Carlo methods, factor analysis, clustering analysis, principle component analysis and Bayesian inference.

·         Professional working experience in Machine Learning algorithms such as LDA, linear regression, logistic regression, GLM, SVM, Naive Bayes, Random Forests, Decision Trees, Clustering, neural networks and Principle Component Analysis.

·         Working knowledge on Recommender Systems and Feature Creation, Validation using ROC plot and K- fold cross validation.

·         Professional working experience of using programming languages and tools such as Python, Hive, Spark, Java, PHP and PL/SQL.

·         Working experienced of statistical analysis using R, SAS (STAT, macros, EM), SPSS, Matlab and Excel.

·         Hands on experience of Data Science libraries in Python such as Pandas, Numpy, SciPy, scikit-learn, Matplotlib, Seaborn, Beautiful Soup, Orange, Rpy2, LibSVM, neurolab, NLTK.

·         Familiar with packages in R such as ggplot2, Caret, Dplyr, Tidyr, Wordcloud, Stringr, e1071, MASS, Rjson, Plyr, FactoMineR, MDP.

·         Working knowledge of NLP based deep learning models in Python 3.

·         Working experience in RDBMS such as SQL Server 2012/2008 and Oracle 11g.

·         Extensive experience of Hadoop, Hive and NoSQL databases such as MongoDB, Cassandra and HBase. 

·         Experience in data visualizations using Python, R, D3.js and Tableau 9.4/9.2.

·         Familiar with conducting GAP analysis, User Acceptance Testing (UAT), SWOT analysis, cost benefit analysis and ROI analysis.

·         Deep understanding of Software Development Life Cycle (SDLC) as well as Agile/Scrum methodology to accelerate Software Development iteration.

·         Experience with version control tool- Git.

·         Extensive experience in handling multiple tasks to meet deadlines and creating deliverables in fast-paced environments and interacting with business and end users.

CERTIFICATIONS:

·         SAS Certified  Advanced Programmer for SAS 9.4                                                                                                             

·         SAS Certified  Base Programmer for SAS 9.4

TECHNICAL SKILLS:

Hadoop Ecosystem	Hadoop2.X, Spark1.6+,Hive2.1, Hbase1.0+
Languages	Python2.7/3, R-3, PL/SQL, SAS 9.4, Hive, Pig, Java, PHP
Packages	Pandas, Numpy, Scikit-learn, Beautiful Soup, GGPLOT2, caret, dplyr, tidyr, wordcloud, stringr, e1071, MASS, rjson, plyr, FactoMineR, seaborn, matplotlib, MDP, Orange, Rpy2, LibSVM, neurolab, NLTK.
Machine Learning	LDA, Naive Bayes, Decision trees, Regression models, Neural Networks, SVM, XG Boost, SVM, random forests, bagging, gradient boosting machines, k-means
Databases	MySQL_5.X, Oracle_11g,SQL_Server2012/2008, MongoDB3.2, HBase 1.0+, Cassandra 3.0
Business Analysis	Requirements Engineering, Business Process Modeling & Improvement, Gap analysis, Cause and Effect Analysis, UI Design, UML Modeling, User Acceptance Testing (UAT), RACI Chart, Financial Modeling
Scripting Language	UNIX Shell, HTML, XML, CSS, JSP, SQL, Markdown
Data Analysis/Visualization	Tableau 9.4/9.2 , Matplotlib, D3.js, Rshiny
Documentation/Modeling Tools	MS Office 2010, MS Project, MS Visio, Rational Rose, Excel (Pivot, Tables, Lookups) Share Point, Rational Requisite Pro, MS Word, PowerPoint, Outlook
Version Control	Git, TFVC
Operating Systems	Linux, Ubuntu, Mac OS, CentOS, Windows

 

 

Professional Experience:

Client: TIAA -- New York, NY                                                        Mar’16 – Till Date

Project Name: Customer Retention

Role: Data Scientist

Project Summary:

TIAA is the leading retirement provider for people who work in the academic, research, medical and cultural fields. 

The project focused on customer retention by addressing client needs in near real-time to predict unsatisfied customers earlier using text analysis and machine learning algorithms in the large structured and unstructured data sources like front-line customer interactions system, Twitter, customer call center etc. Our team also collaborated with business team to apply more customized solutions.

Responsibility:

·         Designed and developed Use Case, Activity Diagrams, Sequence Diagrams, OOD (Object Oriented Design) using UML and Visio with Agile methodology to transform business requirements into analytical goals.

·         Created queries using Spark SQL, Hive, SAS (Proc SQL) and PL/SQL to load large amount of data from MongoDB and SQL Server into HDFS to spot data trends.

·         Wrote Hive-QL to retrieve, query and process raw data.

·         Used Spark SQL to perform data cleansing, transformation and filtering such as identifying outliers, missing value and invalid values.

·         Utilized K-means clustering technique to classify unlabeled data.

·         Worked on data pattern recognition, data cleaning as well as data visualizations such as Scatter Plot, Box Plot and Histogram Plot to explore the data using packages  Matplotlib, Seaborn in Python, ggplot in R and SAS.

·         Used LDA, PCA and Factor Analysis to perform dimensional reduction.

·         Modified and applied Machine Learning algorithm such as Neural Networks, SVM, Bagging, Gradient Boosting, K-Means using PySpark and MLlib to detect target customers.

·         Worked on customer segmentation based on the similarities of the customers using an unsupervised learning technique – cluster analysis.

·         Used Pandas, Numpy, Scipy, Scikit-learn, NLTK in Python for scientific computing and data analysis.

·         Applied cross validation to evaluate and compare the performance among different models. Validated the machine learning classifiers using ROC Curves and Lift Charts.

·         Configured Spark Streaming with Kafka to clean and aggregate real time data.

·         Involved in Text Analytics such as analyzing text, language syntax, structure and semantics.

·         Generated weekly and monthly reports and maintained, manipulated data using SAS macro, Tableau and D3.js. 

·         Involved in using Sqoop to load historical data from SQL Server into HDFS.

·         Used Git for version control.

Environment:

Python 3/2.7, R 3, SAS 9.4, HDFS, MongoDB 3.2, Hadoop, Hive, Linux, Spark, Kafka ,Tableau 9.4, D3.js, SQL Server 2012, Spark SQL, PL/SQL, UML, Git

 

Client: Aetna Group, Hartford, CT                                                          Mar’2014 – Feb’2016

Role: Data Scientist

Project Name: Real-time Fraud Detection

Project Summary:

Aetna Group is an American worldwide health services organization. Its insurance subsidiaries involve in medical, dental, disability, life and accident insurance and related products and services.

The goal of this project is to detect, deter and ultimately end healthcare fraud and abuse. In order to achieve this goal, our team worked on real-time data process and developed fraud detection model using Machine Learning algorithms by analyzing historical claims data, internal fraud, electronic transactions fraud and external fraud. After the model was fine-tuned, we applied this model on real time data to help Aetna detecting fraud more efficiently.

Responsibility:

• Conducted comprehensive analysis and evaluations of business needs; Provided analytical support for policy; Engineered financial, operational and reputational impacts and influence decisions for different models. 
• Retrieved, manipulated, analyzed, aggregated and performed ETL through billions of records of claim data from databases like RDBMS and Hadoop cluster using SAS (Proc SQL), PL/SQL, SparkSQL, Sqoop and Flume.
• Used Matplotlib, Seaborn in Python to visualize the data and performed featuring engineering such as detecting outliers, missing value and interpreting variables.

·         Worked on transformation and dimension reduction of the dataset using PCA and Factor Analysis.

·         Developed, validated and executed machine learning algorithms including Naive Bayes, Decision trees, Regression models, SVM, XG Boost to identify different kinds of fraud and reporting tools that answer applied research and business questions for internal and external clients.

·         Implemented models like Linear Regression, Lasso Regression, Ridge Regression, Elastic Net, Random Forest and Neural Network to provide predictions to help reducing the rate of frauds.

·         Experienced in using Pandas, Numpy, SciPy, Scikit-learn to develop various machine learning algorithms.

·         Used PySpark, MLlib to evaluate different models like F-Score, Precision, Recall, and A/B testing.

·         Fine-tuned the developed algorithms using regularization term to avoid overfitting.

·         Configured Kafka with Spark Streaming API to fetch near real time data from multiply source such as web log.

·         Analyzed real time data using Spark Streaming and Spark core with MLlib.

·         Used the final machine learning model to detect fraud of real time data.

·         Extensively involved in data visualization using D3.js and Tableau.

Environment:

Python 3/2.7, R 3, SAS 9.4, HBase 1.0+, Kafka, HDFS, Hadoop, Hive, Linux, Spark, Tableau 9.2, D3.js, SQL Server 2012, Excel, Spark SQL.

 

Client: Bank of China, Shanghai, China                                                   Nov’2013 – Aug’2014 Project Name: Fraud Detection

 

 

Role: Data Scientist

Project Summary:

Bank of China was the second largest bank in mainland China.  The bank offers a full range of services that cater to both middle-class individuals and business-oriented individuals.

The goal of the project is to collaborate with analytics team to detect fraudulent activities to determine if an ongoing transaction is legitimate or not.

Responsibility:

·         Created new features based on information from million transaction records and training models using Machine-Learning techniques such as Gradient Boosting Tree and Deep Learning.

·         Analyzed and determined a cutoff point for accepting/ declining transactions to minimize fraud losses and increase customer experience by using various machine learning algorithms such as Logistic Regression, Classification, Random Forests and Clustering in SAS, R and Python.

·         Used Pandas, Numpy, Seaborn, Scipy, Matplotlib, Scikit-learn, NLTK in Python for implementing various machine learning algorithms.

·         Used SAS, SQL, Oracle, Teradata and MS Office analysis tools to complete analysis requirements. Created SAS data sets by extracting data from Oracle database and flat files

·         Used Proc SQL, Proc Import, SAS Data Step to clean, validate and manipulate data. 

·         Performed updating data by weekly and monthly; maintained, manipulated the data for database management. Used the SAS Macro and Excel Macro for the monthly production. 

·         Experienced in SQL queries to retrieve and validate data, prepared for data mapping document. 

·         Worked on RDBMS like MySQL and NoSQL databases like MongoDB.

·         Used the Agile Scrum methodology to build the different phases of software development life cycle. 

Environment:

SAS9.4, Base SAS, SAS Macros SAS Graph, SAS Access, SAS STAT, SAS ODS, SAS SQL, SAS/ETL, SAS/Stat, SAS ENTERPRISE Miner, Python, PL/SQL, Oracle 9i, Hadoop, MongoDB

Client: Alibaba Group, Hangzhou, China                                                   Jan’2013- Oct’2013                                                              

Project Name: Customer Behavior Analysis

Role: Data Analyst                                                                                       

Project Summary:

Alibaba Group Holding Limited is a Chinese e-commerce company that provides  consumer-to-consumer, business-to-consumer and business-to-business sales services via web portals. It also provides electronic payment services, a shopping search engine and data-centric cloud computing services.

The project was designed to improve their overall health, growth, and profitability.

Responsibility:

·         Analyzed online user behavior, conversion data and customer journeys, funnel analysis and multi-channel attribution.

·         Worked on business forecasting, segmentation analysis and data mining.

·         Involved in the development of Data Warehouse for personal lines property and casualty insurance.

·         Generated graphs and reports using ggplot in RStudio for analyzing models.

·         Developed and implemented R and Shiny for business forecasting.

·         Developed predictive models using Decision Tree, Random Forest and Naïve Bayes.

·         Used available data sources to deep dive and troubleshoot campaign performance issues.

Environment: MySQL5.5, R 3, caret, Shiny

Client: Ping An Insurance (Group) Company of China, Ltd    

Role: Data Analyst                                                                                       Jan’2011- Dec’2012

Project Name: Claims data analysis

Project Summary:

Ping An Insurance (Group) Company of China, Ltd. is a holding company whose subsidiaries mainly deal with insurance, banking, and financial services.

The project was designed to improve the Disease Management and Outcomes in order to support various care plan.

Responsibilities:

·         Provided analytical support for the Claims, Ancillary, and Medical Management.

·         Performed Data Mapping and Logical Data Modeling; Created class diagrams, ER diagrams

·         Cleaned data by analyzing and eliminating duplicate and inaccurate data using PROC FREQ, PROC MEAN, PROC UNIVARIATE, PROC RANK, and macros in SAS; Used SQL queries to filter data.

·         Converted various SQL statements into stored procedures thereby reducing the number of database accesses.

·         Worked with Quality Control Teams to develop Test Plan and Test Cases.

·         Designed and implemented basic SQL queries for Data Report and Data Validation.

·         Developed user manuals and provided orientation and training to end users for all modified and new systems.

Environment: Base SAS, SAS/Access, SAS/Stat, SAS/Graph, SAS/SQL, SAS/ODS, SAS DI Studio, SAS/Macros, MS Excel, MS Word, PowerPoint, Oracle 9g, DB2, MS Excel, UNIX, SAS ENTERPRISE Miner, SAS EBI (Enterprise Business Intelligence) 9.4

Education:

Master of Science in Information Systems

Bachelor of Science in Information Systems