Garrett McKenzie

My professional experience in relevant fields is limited; nevertheless, my educational experience, personal experience, and work in other fields makes me well suited for team oriented tasks involving programming, machine learning, or data analysis. Further, I love learning, and am fully prepared to put in extra hours to fully comprehend a vital piece for a project.

Relevant Projects:

IMBD Analysis

In this project, my partner and I analyzed a set of 10,000 IMDB movie samples with information for each sample such as rating and genre. The purpose of this project was to understand if there was a correlation between movie genre and mean meta-score of movies in that genre. For this project, extensive data cleaning and preprocessing was needed before it could be successfully visualized. After this ANOVA and Pairwise T-test, with appropriate Bonferroni adjustment, were used to reject the null hypothesis in favor of the hypothesis that at least one genre has a significantly different mean meta-score. Further, it was concluded that the animation genre has a significantly higher mean meta-score, and the horror genre has a significantly lower mean meta-score.

Analysis of Professional Fighting Game Players

In this project, my partner and I analyzed data we scrapped from a website with rankings for professional e-sports players. The purpose of the project was to develop an regression algorithm to predict a player's ranking given certain features about them. After data scrapping, preprocessing, model selection and hyperparameter tuning, we successfully developed a regressive decision tree algorithm that predicted a player's DF score, which is how they are ranked, with a MAE on the test set of 357 points. This MAE was considered acceptable as it beat the prior, and as many relevant players had DF scores greater than 1000.