My name is Sam Kinghorn, and I'm passionate about statistics and programming. I have a Master's in Statistics, which has equipped me with the tools and techniques needed to understand data and extract valuable insights. The culmination of my degree was a presentation of my thesis on merging neural networks with survival data to automate the model selection process and increase prediction accuracy. I'm interested in all aspects of the data workflow, from data collection and storage to exploration, experimentation, and prediction. I'm a highly motivated person with a desire to produce high quality ouput and learn more everyday.
Master of Science in Statistics
Ball State University
Graduated: May 2023
GPA: 4.00/4.00
Thesis: Weibull Neural Network Survival Models with Frailty
Bachelor of Science in Accounting
Ball State University
Graduated: May 2021
GPA: 4.00/4.00, summa cum laude
The following are personal projects I've worked on as part of my portfolio. In these projects, I seek to uncover insights in data through visualizations and descriptive statistics as well as create models to determine relationships among variables. Contact me to learn more.
This project looks at hurricane data in R to determine if the femininity level of hurricane name is associated with more deaths. I perform an exploratory data analysis as well as fit poisson and negative binomial regression models to the data. I conclude that there is no association between femininity level of hurricane name and hurricane deaths.
In this project I analyze data from a randomized clinical trial on the effectiveness of two treatments for smoking cessation. I perform a survival analysis where the goal is to determine which intervention extends the time to relapse for smoking cigarettes. I visualize the data and use the Cox proportional hazards model to determine which treatment is most effective.
I utilize PyTorch to create a custom Weibull neural network model to predict customer churn. I developed this model as part of my master's thesis and extend this idea to frailty models. These models have wide application in health settings as well as business settings (e.g. churn, credit risk, etc.). These ideas can be extended to other parametric models.
In this project I scrape IMDB user reviews for the movie Tenet to get a sense of the broad sentiment surrounding the film. I utlize Selenium and BeautifulSoup to extract and parse the HTML.
