Data Science

Ryan Liao

Relevant Skillsets

Working with and manipulating data frames in Python, Java, Dask, Spark, & SQL

Converting cleaned data into visuals using HTML, Javascript & D3.

Using R and Stata to interpret public policy data.

Extensive experience in Microsoft Excel/Google Sheets. Relevant Coursework Experience

Language Familiarity, Data Structures & Algorithms

- Programmed advanced Python & Java using object oriented programming with different data structures to create different scenarios using stacks, queues, recursion, etc.

- Applied calculus, probability theory, and other statistical concepts to solve practical problems.

- Examined the efficiency of various algorithms (binary search trees, hash tables, BFS, DFS, etc.) and learned graph theory with adjacency matrices.

Large Scale Data Management & Machine Learning

- Learned basic to advanced SQL via SQLite and PostgreSQL querying.

- Discussed conceptual data management topics such as conceptual design, transactions, storage, etc.

- Extracting large data sets from AWS and used Dask and Spark to query and manipulate them.

- Applied text mining and sentiment analysis fundamentals to public policy and consumer review text data.

- Trained features with various regression and classification algorithms for commercial recommender systems. Project Experience

Clothing Size Recommender System Algorithm

- Fitted a machine learning algorithm to predict a user’s most comfortable size based on their clothing reviews.

- Sourced dataset from RentTheRunway and identified predictive features and suitable classification methods.

- Tested each algorithm using various summary statistics and eventually settled on XGB regression with mean-squared error as the best performing solution.

- Condensed methodology and findings into a brief academic paper which also includes analysis of similar studies and algorithms made in the past.

Fully Interactive Heads Up Poker Game & Rock Paper Scissors Game

- Created a functional one on one poker game and rock paper scissors game using Python and Easygui.

- The poker game was programmed to be a fully customizable cash game in which stack sizes and blind sizes could be changed to the players’ choosing.

- Relevant statistics for each game such as user ID, balance and win rate were stored in an SQL database and exported to a CSV after each session.

NFL Regular Season Team Unit Success Correlation with Team Playoff Success

- Did an extensive analysis using data frame wrangling, visualizations, and linear regressions to determine how well regular season offensive or defensive strength would predict playoff success.

- Utilized modules such as Pandas, Seaborn, and Matplotlib as well as the Pro Football Reference API.

Detailed Analysis on China’s Past and Future Soft Power Influence

- Identified different factors towards China’s soft power expansion in recent decades in order to assess which parts of the world they have most targeted.

- Scraped datasets from various sources regarding FDI, UN voting, HDI, Trade, GDP, etc. to formulate a model which takes every variable into account.

- Predicted China’s future soft power targets using regression and visualized their influence using various geoplots in R ggplot.

- Compiled all findings into a 30 page research paper outlining every variable in detail as well as the overall methodology.

Senior at the University of California at San Diego

Data Science Major / Political Science Data Analytics Minor

Available for Data Science, Analysis, & Engineering Related Positions starting April 2023. Phone: 469-***-****

Emails: ******@****.***

**************@*****.***

U.S. Citizen

Contact this candidate