Home
Other sites:
GitHub: https://github.com/rufuspetrie
LeetCode: https://leetcode.com/big_rufus/
HackerRank: https://www.hackerrank.com/rufuspetrie
Some interesting projects:
Indian Buffet Process Paper: (Paper) (Code)
Wrote paper describing implementation of Indian Buffet Process algorithm
Used NumPy to implement Markov Chain Monte Carlo model with IBP prior
Optimized algorithm with vectorization, just-in-time compilation
Demonstrated algorithm by performing latent feature analysis on simulated images
Analysis of Wisconsin Basketball: (Writeup) (Code)
Investigated large database of information concerning NCAA basketball on Bigquery
Used SQL to query information about schools, programs, players, and game events
Generated statistics and graphics for query results using Python
Wrote paper evaluating history and overall strength of Wisconsin men's basketball program
Analysis of Melbourne Real Estate Market: (Writeup) (Code)
Created real estate dataset by combining web listings, tax records, and geospatial data
Analyzed postcodes by clustering into groups, generating maps with Geopandas
Used Python to statistically analyze and model prices for individual listings
Found that Random Forests could explain over 82% of variation in the data out of sample
Paper about Predicting MLB Attendances and Prices: (Paper) (Code)
Wrote paper analyzing MLB attendances and resale prices of MLB tickets on secondary markets
Used Scikit-learn to build predictive models for ticket prices and attendances using Stubhub data
Showed that the timing of ticket sales makes little difference when predicting ticket prices
Demonstrated significant differences in behavior between frequent and casual sellers on Stubhub
Paper about Healthcare and Development: (Paper) (Code)
Wrote paper analyzing elasticity of healthcare spending in a development setting
Used R and Stata to clean survey data and generate appropriate econometric models
Found no immediate statistical relationship between pregnancies and healthcare spending
Found that households may delay healthcare spending 1-2 years following childbirth
Paper about Efficacy of Death Penalty: (Paper) (Code)
Wrote paper attempting to determine whether the death penalty impacts crime rates
Used Stata to build models evaluating relationships between death penalty and various crime rates
Found mixed results but overall significant correlation between death penalty and crime rates
Discussed overall feasibility, pitfalls of studying association of the death penalty with various crimes
Algorithmic Trading Project: (Code)
Implemented automated trading algorithm using Python on the Quantopian platform
Selected securities to trade based off of trailing returns and CVaR
Hedged positions by calculating market correlation of portfolio and shorting index funds
Record modest returns and minimal market correlation over duration of project
Neural Network for Handwriting Recognition: (Code)
Implemented Neural Network from scratch in Python
Trained Neural Network to classify handwritten numbers from the MNIST data
Recorded over 98% classification accuracy out of sample on the MNIST data
Correctly classified all of my own handwritten numbers
Completed Project Euler Problems: (Code)
Some solutions to an old Project Euler competition on Hackerrank
Used Python to solve computationally intensive number theory problems
Scored in top 2% while active in competition
Stanford Online Certificates:
Algorithm Design and Analysis: https://bit.ly/3dqjTfU
Part 2: https://bit.ly/31uhYVm
Statistical Learning: https://bit.ly/3GnmBPH
Hackerrank Certificates:
Problem Solving: (Basic)
Python: (Basic)
C++: (Basic)
SQL: (Basic) (Intermediate)