Data Analyst Marketing Manager

Ted Cai ************@*****.*** C. 520-***-**** San Mateo, CA

https://www.linkedin.com/in/haomeng-cai/ https://github.com/HaomengCai SUMMARY

● Data Analyst with 3 years of work experience in finance & high tech operation domains

● Expertise and hands-on experience in SQL, big data analytics, data visualization, machine learning, and experimentation TECHNICAL SKILLS

● Tableau SQL R Python SAS Hadoop Hive Keras PowerPoint Advanced Analysis in Excel AWS

● Data Visualization Quantitative Methods Data Mining Regression Analysis Machine Learning Exponential Smoothing EDUCATION

M.S. Business Analytics California State University, East Bay Hayward, CA Dec. 2019 Core Coursework: Database Management & Applications (A), Data Analytics (A-), Big Data Tech. & Apps (A-) B.A. Economics, Minor: Business Administration University of Arizona Tucson, AZ Dec. 2016 WORK EXPERIENCE

Data Analyst, Facebook Inc.(via Spectraforce), Menlo Park, CA Feb. 2020 – Now

● Write SQL queries to extract data from Facebook internal Presto, develop strategic metrics to providing valuable insights for the team to improve the finance operational efficiency

● Estimate operational performance by analyzing millions of rows of data, use normal distribution, linear regression, etc. statistical methods to interpret business performance and risk factors and predict procurement activities

● Convert analytical findings into 20+ live data dashboards in Tableau and internal tools, effectively act upon of procurement process for internal stakeholders and suppliers, impacted 600 cost centers and ~$5 million operational cost Graduate Researcher, California State University- East Bay, Hayward, CA Aug. 2018 – Dec. 2019

● Utilized Tableau to visualize student employment placement data via pivot table, pie charts and provided data for marketing

● Conducted A/B tests on Facebook Ads for different online ad campaigns, interpreted statistical significance of experiment results, and proposed final recommendations on ad variations and audience selections, which improved CTR by 17% and

$32k expected incremental revenue per quarter

● Built time series, seasonality model in Python to forecast shipping expenses for marketing materials for university advertisement, presented results to marketing manager to make smarter budget decision Data Analyst, Lusida Rubber Products Inc., Los Angeles, CA Jan. 2017 – Jan. 2018 Global supplier of custom rubber and plastic components for the automotive, industrial, HVAC, and appliance industries

● Estimated prices of materials for customer projects based on domain knowledge and historical data analysis

● Wrote SQL queries to analyze the quantity and price of materials from 25+ factories overseas using MySQL Database and data analysis tools in advanced Excel tool to provide quotes for manager

● Created Salesforce reports to estimate the best shipping methods (airfreight, overseas carrier shipping, etc.) and saved the company ~ $130K in implementation costs per month

● Maintained multiple tables in MySQL to record transactions, purchase orders, and inventories; collected over 1K related data records per day and generated 10+ reports for production metrics PROJECTS

E-invoicing Live Dashboard in Tableau Server, Facebook Feb. – Apr. 2020

● Cross joined Facebook data sources from the internal Oracle and Presto database, extracted performance-related dimensions and features, maintained 1,000,000 rows of live data daily for over 20 operation related tables

● Broke-down the invoicing source by multiple dimensions and created charts in Tableau dashboard, ranked the supplier invoice source types and list performance changes quarterly by business unit

● Published over 20 dashboards in Tableau, providing comparisons and insights for over 130K Facebook employees and 600 cost centers based on dashboard data visualization, improved the operational efficiency by 30% and satisfaction of suppliers and internal stakeholders

Yelp Customer Review Data Challenge for LSTM Modeling and Prediction Analysis Jun. – Aug. 2019

● Data Preprocessing : Sampled 10% data, 250k rows from Yelp challenge raw dataset for LSTM sentiment analysis to predict positive, negative, or moderate reviews; split the dataset into validation data and training data by portion 20%: 80%; visualized the distribution of each class to prevent oversampling; and converted to one-hot encoding and transformed sentences into word vectors for model training and predictions

● Modeling : Built LSTM layer and Bidirectional LSTM layer, used Adam optimizer and categorical cross-entropy as loss function, and achieved 85.9% and 86.1% accuracy for LSTM model and Bidirectional LSTM model on validation datasets

Contact this candidate