Steven Ta Email & Phone Number

San Francisco Bay Area

• Feature engineer restaurant data taxonomy to be used for machine learning job recommendation system.• Normalize third party data on 50,000+ restaurants using Python Pandas and NumPy.• Create daily reporting dashboards for business intelligence team using Redshift, Superset and Periscope.• Develop analytics reports with summaries and recommendations for Product, Marketing, and Sales teams.• Monitor page level metrics and benchmarking success through Google Analytics.

San Francisco Bay Area

Galvanize Capstone ProjectContent Based NLP Hotel Recommender Project Link: https://github.com/steventa87/Hotel_Reccomendation• Problem: Ratings are heavily relied upon in today’s ecommerce world. Other than ratings, a customer would read a couple of reviews and then make their purchase decision. This leaves thousands of reviews about other customer experiences wasted.• Objective: Build a web application that would utilize customer experiences to recommend a hotel based on anexperience a user wants and what type of personality they have.• Data: Trip Advisor – Analyzed over 600 hotels and over 100k reviews in U.S cities such as San Francisco, Las Vegas, and New York City. User content were without star or numeric ratings.• Tools: TF-IDF, Count Vectorizer, Cosine similarity, Vader Sentiment Analysis, Latent Dirichlet Allocation, LDAvis, K-Means Clustering, Flask.• Results: Completed web application where a user can input an experience or personality trait and web app will produce top 3 recommendations. Model splits recommended hotel’s reviews by sentiment of positive, negative, neutral and provides unsupervised topic modeling. Topic model clusters 5 main categories to give insight as to where the positive and negative features are truly from.

• Ran SQL queries to do weekly audit of courier pay rates to maximize profits and minimize courier wait times.• A/B tested value of time sensitive flash promotions, seasonal trends, and weather impact.• Managed transition of GPS-based fulfillment network from manual to an automated system.• Piloted the Courier Program to East Bay and Sacramento courier teams and trained internal employees.• Ran up to 3 onboarding sessions of interviewing, screening, and hiring a day to meet demand of 50 active drivers per week.

San Francisco Bay Area

• Established the San Francisco Courier Program and maintained 500 active drivers with an initial growth of over 500% to fleet in the program’s first quarter.• Navigated up to 300 couriers daily from all points while maintaining highest pay and most efficient route between San Francisco and East Bay.• Managed day to day response rate, refund rate, and bonus rate on Salesforce CRM.• Moderated customer feedback review system and incentivized highest performers with bonuses.