Data scientist with 5+ years of engineering work experience. An authority on defendable, data-driven decision making after years of working under FDA scrutiny.
Strives to bring care and creativity to every problem. Known for fast learning, self-motivated research, and unrelenting quality.
A selection of projects developed as part of Data Science coursework at UC Berkeley
Click on a project name to view documents
Project experience performing statistical analyses, developing models for prediction, categorization, and language processing, and presenting findings to enable decisions.
Beating the Game: Predicting Board Game Success (Machine Learning) This notebook uses board game data to understand what makes a successful design. It explains data filtering and feature selection, then creates kNN, Naive Bayes, and Linear Regression models, with insights extracted from each. Finally, it compares games from different decades to demonstrate how trends have changed. (click here to view summary slides)
Sentiment Analysis of Political Statements
Using Cross-Field Training Corpora (Natural Language Processing) This project extracts
statements regarding healthcare from congresspeoples' .gov websites, then performs NLP analysis to
predict how those politicians would vote in matters related to the ACA. After employing various known
language sets for training, the project compares predictions to actual voting and ultimately
uncovers interesting nuances specific to
processing political language.
(click here to view summary slides)
Ford GoBike Business Analysis (SQL) This notebook presents an analysis of Ford GoBike use in San Francisco, CA. It frames the business case, pulling and displaying relevant information from a Google BigQuery database. Finally, it makes a recommendation for offering a particular coupon to users.
NBA Statistics Analysis (Tableau) This exploratory data analysis uses Tableau to display NBA player statistics in multiple forms, allowing the reader to visually evaluate a few basic hypotheses.
The Danger of an Over-personalized Insurance Industry This paper discusses the existing ways in which personal data has entered the insurance industry: demographics, driving metrics, fitness data, etc. It then considers other sources of data that would be useful for outcome prediction (social media, DNA) and explores whether the current legal regulations are sufficient to protect individual privacy. (click here to view summary slides)
Data Science Skills
Software & Tools
Data Scientist
Mechanical Engineer
Coursework included data engineering, statistical methods, experiments and causality, machine learning, natural language processing, data visualization, and data ethics
Coursework included engineering design, kinematics and robotics, control theory, learning systems, and programming methods