Customer Segmentation
A retail analytics project using the UCI Online Retail dataset (500K+ transactions) to drive business insights. Cleaned and processed raw sales data, achieving >90% data quality retention. Engineered advanced customer features (RFM, frequency, diversity), performed exploratory analysis to highlight sales peaks and top products, and conducted correlation analysis to optimize feature selection. Segmented 4,000+ UK customers via K-means and GMM, creating actionable profiles (e.g., high-value, loyal), and built predictive models (Random Forest, XGBoost) that improved CLV prediction accuracy by 30%. Delivered clear recommendations for targeted marketing and retention strategies.
View on GitHub
Movie Suggestions
For my HarvardX Data Science capstone project, I developed a movie recommendation system using the MovieLens 10M dataset, with the goal of accurately predicting user ratings. My approach involved building a series of models in R, starting with a simple average and then iteratively incorporating movie and user-specific biases to account for patterns in the data. The final, most successful model was a regularized linear model, where I tuned the penalty parameter (lambda) to penalize movies and users with few ratings, thus preventing overfitting and improving predictive accuracy. This final model achieved a Root Mean Squared Error (RMSE) of 0.84, successfully surpassing the project's performance target and demonstrating my ability to wrangle large datasets, implement machine learning algorithms, and perform robust model validation.
View on GitHub
LinkedIn Scraper
This is a tool I developed for my networking during the 10th cohort of the Accenture & ONETEN Scholarship for front-end development. It helps me keep up with the very large amount of contact information the whole cohort overall (two other programs including Digital Marketing and Business Analytics) and maintain a clean list of the available linkedin profiles found on other students profile cards. In the future I will probably be adapting this to other cohorts since it is fairly modular as well as flexible given the tag only scrape or JavaScript activities scrape portions of its sweep.
View on GitHub
Bicycle Tracker
This is another side quest project for one of my hobbies. I wanted to create a bicycle tracking web app that uses GPS to display the user's real-time location, and speed. The app dynamically updates the map and info card every two seconds, fetching nearby cross streets via OpenStreetMap's API. By allowing users to toggle between light and dark themes safety at night is also taken into consideration due to other maps having brightly colored ads. The app ensures high accuracy by leveraging GPS, Wi-Fi, and cellular networks, providing a responsive and user-friendly interface.
View on GitHub
MoreAvgWeather
This is a side quest project I started while taking a break from coding lessons to focus on mathematics. I've been upskilling my programming through edX and my local library and found a probability course that will help me dissect more complicated kinds of equations that get used in algorithms. This project uses the known phase changes of water, a users altitude, and a collection of basic weather station readings from four different weather service providers to try to predict the probability of rain as precipitation in a given part of downtown Seattle WA.
View on GitHub
Weather App
This is a capstone project for CS50W to explore JavaScript, Python, CSS, Django and other web development frameworks. This application displays the current weather and precipitation data to a user who must register to access the site.
View on GitHub
CodeCampFriendly
This is my final CS50P Python project. It scrapes an API to aggregate all of the coding bootcamp friendly job listings for Seattle into one place without having to scroll through advertisements.
View on GitHub
Personal Portfolio
This is a project from CS50x that I put together at the end of the course. Its a portfolio website that contains my resume information, class certificates, and a page for my favorite projects that I've worked on so far.
View on GitHub