Projects
Here is a sample of the projects I have worked on in the past. These mainly include the projects I have done in class and some personal ones.
Analyzing Factors of Professional Sport Teams
My first major project in my graduate school studies involved heavy linear regression and data visualization. The goal was able to find potential factors that lead to the professional sports teams' success. This was all done in R and was my first exposure to cross-sectional data and large data sets that with observations that exceeded 1,000.
Inelasticity of Gas Prices and Forecast
My first exposure to time series data and its analysis of it came when I was tasked to build a model that could forecast and predict potential gas price spikes from the war in Ukraine. I performed stationary tests using EViews to help determine which data sets would need to be transformed in order to not skew the results of my analysis.
Buxton Store Opening
My largest group project had me collaborate with a team of seven to use a large data set sourced from a real company to help give a real-world experience. The company had the goal of being able to predict which potential store location would bring in the most revenue. This was my first time experiencing a new data analytics program named Gauss. The biggest experience I gained from this was managing the teams' effort and resources properly in order to meet the deadlines and have a proper presentation to show the company.
Buxton Store Opening
My largest group project had me collaborate with a team of seven to use a large data set sourced from a real company to help give a real-world experience. The company had the goal of being able to predict which potential store location would bring in the most revenue. This was my first time experiencing using Stata which is a popular choice for data analytics. The biggest experience I gained from this was managing the teams' efforts and resources properly in order to meet the deadlines and have a proper presentation to show the company.
Replication Project of Student Beauty Study
I was also tasked with being able to replicate the research of a foreign professor. The research I replicated was done by a Swedish professor who ran regression models on data collected from the students at his university. The purpose was to help showcase potential links with student beauty and their grades in certain classes. I expanded on his research by considering different variable relationships and potentially using different models. This was all done in Gauss which was a different programming language than the original professor use. The biggest challenge was both the language barrier with the data set and the original programming language conducted being different than mine.
Predicting Video Game Success Through Machine Learning
A very large number of popularized machine learning methods were taught in one semester of my graduate studies. This was a fast-paced course that covered a large variety of machine learning and deep learning methods. I had to collect data through web scrapping video game data and clean via Python to ensure the estimation models were accurate. The main goal was to be able to build a model that could accurately predict the potential success video game projects could have with certain input variables. A large variety of machine learning methods were considered such as linear regressions with splines, bagged linear regressions, regression trees, AdaBoosting, and Stochatic Gradient Boosting. Other deep learning methods were also used such as neural networks that was popularized from programs such as ChatGPT.
Pokemon TCG Prediction Models
In order to keep being ambitious and passionate about my learning and another outlet for showcasing my technical knowledge and skills, I created a personal project for myself. The purpose is to connect my own interest to my schooling to continue my exposure to data analytics methods. With the methodology I learned in the classroom, I was able to create a data set filled with Pokemon Trading Card Game observations. The purpose was to be able to create all sorts of models that have the goal of predicting the price of certain Pokemon cards and potentially seeing what the future market price will be. The market of this niche hobby is very volatile and often thought of as hard to predict. The models I have created showcase that it may be hard to predict, but there are obvious historical trends that could potentially explain why certain cards may hold their value better than others.
