YIMENG ZHAO

Email: aczv1r@r.postjobfree.com Phone: (512) 713 – 9628 Github: https://github.com/adv34715 EDUCATION

M.S, Information Science, University of Texas at Austin; Austin, TX August 2015 - May 2017 Specialization: Data Science, Data Mining, Database Management, Web Development. GPA: 3.9 B.E, Electronic Technology, Hefei University of Technology; China August 2011 - June 2015 Course Algorithms and Data Structure, Pattern Recognition. EXPERIENCE

Data Analysis Intern, IC2 Institute, Austin, TX; June 2016 - December 2016

• Designed a schema and built a longitudinal database for Austin's entrepreneurs and startups using MySQL Workbench.

• Collected and analyzed data to verify universities academic development stimulate Austin's economic development.

• Collaborated with team members on research and presented findings to different audience. PROJECTS

Research on Rationales Effect on Crowdsourcing Workers January 2017 - Present

• Designed crowdsourcing tasks for crowdsourcing workers to compare work quality with/without rationales provided.

• Designed webpage prototypes and build webpages for crowdsourcing tasks using JavaScript/HTML/CSS which can collect workers’ behavior data, such as mouse strokes and draw operations.

• Evaluate workers’ work quality by their accuracy and behavior data and analyze rationales effects on workers. Student Paring Systems, PenPal Schools, Austin, TX; January 2017 - Present

• Extract and aggregate useful/related data from PostgreSQL database using psycopg2 and store data into csv files.

• Train classifiers with data from cvs files to classify students into active and inactive ones with cross-validation method.

• Provide students recommendations for choosing group members and new online courses. Crowdsourcing Workflow for Generating Human Data September 2016 - December 2016

• Designed a program that users can set filters to select historical medical image based on data derived from MongoDB.

• Designed an algorithm to improve the data quality and tested the algorithms feasibility and stability. Bank Telemarketing Success Prediction January 2016 - May 2016

• Selected 20 features from more than 50 features and applied SVM, LR, DT, RF classification models to predict a bank's telemarketing success on about 50, 000 records.

• Used cross-fold test method to evaluate the four classification models and to find the best parameters for each model, as well as explored how 20 feature (numerical and categorical) of telemarketing affect the success of telemarketing. SKILLS&TOOLS

Languages Python, MySQL, C++, JavaScript, HTML, CSS, Bootstrap, Unix commands. Databases MongoDB ( pymongo), PostgreSQL (psycopg2), MySQL Workbench. Softwares/ Tools Jupyter Notebook, Github, Amazon Mechanical Turk, AWS, Photoshop, Overleaf, Qlik Sense, Microsoft Office Suite.

Design Dashboards, Posters, Infographics, Slides.