YAN LI

**** ***** ***** **, ******, TX ***** ●972-***-**** ●adm15k@r.postjobfree.com

PROFILE

Python/SQL Programmer and Data Analyst with strong Statistics/Econometric background and 2 years of experience using data cleaning, data visualizing and data mining algorithms to solve challenging business problems. Have customer behavior Research skills in large sets of structured and unstructured data, Modeling/Reporting, Seeking fulltime position in the field of Data Analyst and Business Data Analytics

TECHICAL SKILLS

SQL • SAS(Advanced Certification) • Python • Tableau • Excel • SPSS • IBM db2 Server • R EXPERIENCE

Ddata Analyst Dallas, TX

Elink-Elite Solution June/2020 – Now

• Extracted mobile phone log data and test result data using SQL and Excel to build daily summary testing report

• Cooperated with internal business leaders to test, establish, improve key performance indicator (KPIs) on ad-hoc needs

• Translated statistical findings into business insights and presented to senior management team on a weekly basis using Tableau

• Worked closely with marketing team to develop our advanced clients analysis report on a monthly basis

• Lead overall data management solutions and provided cleaned, well-constructed data input to Tableau visualization dashboards Data Analyst Dallas, TX

TaxFree Shopping Aug/2019 – June/2020

• Extracted customer tax data and validated and edited completed cash flow sheet and balanced sheets daily for accuracy

• Designed, developed, tested, and maintained reports and analyses to drive key business decisions using SQL and Tableau

• Worked with Python and Excel to perform analysis of domain specific source data to obtain key profit drivers

• Researched and discovered predictive patterns using Tableau for data visualization and Python for trends investigation

• Delivered a final report showing cross-functional benefits using Power Point, Excel and Tableau Private Equity Analyst Assistant Internship Beijing, China Tiantu Capital Management Center Dec/2016 –Feb/2017

• Involved with first trail of Haiwan Financing Project and surveyed background of CEO in Haiwan company.

• Gathered and Extracted top 10 outbound tourism Chinese travel companies form Tainyancha financial database using Python

• Conducted data regression analyses of competitor company and tourism industry market distribution

• Visualized and analyzed financial and management reports from Aug. 2014 to Feb. 2015 using Tableau and Excel

• Reported and recommend to senior management team to invest 7 to 9 million with obtaining stock right of 12%-13%. ACADEMIC PROJECTS

Time Series ARIMA model and Cointegration ECM Model in SAS

• Visualized two time series data in last 15 years for testing trend, intercept and seasonality using SAS PROC SGPLOT

• Performed ADF unit root test for stationarity and developed ARIMA model to forecast orange price

• Performed Granger causal test for independence Engle-Granger Residual-Based test for cointegration.

• Created Cointegration ECM model and predicted orange price in next 6 months with 92% accuracy in test data. Multiple Linear Regression and Association Rules in Python

• Visualized and analyzed Stepwise regression significant features: house square feet, number of bedrooms, number of schools, distance of neighbor for target variable house sale price

• Developed multiple linear regression model to predict house sale price in Ames, Iowa

• Evaluated model with MSE and improved model accuracy to 90% in test dataset Predicted American Airline flights delay possibility in DFW airport using Python

• Created a custom, one-versus-all logistic regression classifier using a customizable regularization(L1/L2).

• Adjusted the optimization technique and the value of the regularization term "C" to achieve 78% accuracy.

• Compared the performance of the "best" logistic regression optimization procedure to scikit-learn model. Multi-Layer Perceptron: Analyzed trip types of Walmart consumers. (Neural Network in Python)

• Visualized the distribution of features and explored the imbalance class of features in 10,000 dataset.

• Created a custom implementation of the multi-layer perceptron with adjustable cost function (quadratic and cross entropy), adjustable activation function (ReLU, SiLU, sigmoid) and adjustable number of hidden layers.

• Tuned hyper-parameters of the MLP model with Stratified 10-fold cross validation, improved f1 score to 0.72. Recurrent Neural Network: Classified aggressive and nonaggressive tweets to identify cyber-bullying with Python

• Cleaned, Tokenized and Embedded 20000 tweets text data to integer matrix with Keras API of Python.

• Created RNN model with three different recurrent network architectures (SimpleRNN, LSTM, GRU), adjustable embed size and dropout.

• Visualized the ROC curve of models with 0.82 f1 score and Stratified 10-Fold.

• Performed ADF unit root test for stationarity, Granger causal test for independence of two time series Engle-Granger Residual- Based test for cointegration.

• Created Cointegration ECM model and predicted orange price in next 6 months with 92% accuracy in test data Master of Science from SMU; STEM and OPT certified

Contact this candidate