Mental health clustering analysis

I collect mental health data from over 1800 participants and analyze the data using k-means clustering

LOST transcript analysis

Analysis of transcripts form the hit tv show Lost. Which characters are known for their use of "dude", "aye", "ain't", and "bloody"?

Fiction text analysis

What words define fiction compared to other genres like nonfiction, newspapers, and magazines? I use a simple frequency based approach to answer the question

Professional boxer analysis

In this analysis I use Empirical Bayesian techniques to better estimate a boxer's win rate. This technique is especially effective for boxer's who have a very small number of boxing matches under their belt

Master angler analysis

I scrape 400,000+ trophy fish records from Manitoba and analyze fish catching trends over time, women's involvement in sport fishing, as well as find several hot fish/lake/season combos for catching big fish

Bob Ross IMDb rating predictions

I use several machine learning models to predict IMDb episode ratings for Bob Ross' The Joy of Painting episodes

LOST IMDb episode rating prediction

I use several machine learning models to predict IMDb episode ratings from text descriptions of Lost episodes as well as analyze the defining words of each season

Birthday problem simulation

I solve the famous Birthday Problem via Monte Carlo simulation

Monty Hall simulation

I solve the Monty Hall problem through simulation showing why you should always "switch"

Median split simulations

I show why it is never beneficial to perform a median split (and other splits) on a continuous variable. The main issue is a reduction of statistical power in finding effects (i.e., type 1 errors)

t-test simulations

In this brief simulation, I demonstrate the dangers of violating the equal sample size and equal variance assumptions while using the Student's t-test. The Welch's t-test also shows it's impressive ability to correct for assumption violations

Computing Fibonacci numbers using eigenvectors

I show how to represent the algorithm that generates fibonacci numbers as a matrix, then use a linear alegbra method called eigendecomposition to generate any fibonacci number without having to compute the preceding numbers in the series