Projects
Featured Work
Below are selected projects that demonstrate my technical skills and analytical thinking across machine learning, statistical modeling, and software development.
U-Know-Uno: Probabilistic Game Solver
Live Site: u-know-uno.netlify.app
Repository: github.com/bgwarren1/U-know-Uno
An interactive web application that calculates probabilistically optimal moves for the card game Uno using Monte Carlo simulation and game tree analysis.
Key Features:
- Real-time Optimization: Analyzes game state and computes optimal play strategies in milliseconds
- Probability Visualization: Interactive charts showing win probabilities for each possible move
- Custom Game Logic: Implements full Uno rule set including special cards, stacking, and challenge mechanics
- Responsive Interface: Clean, intuitive UI built for both desktop and mobile gameplay
Technical Implementation:
- Developed custom probability engine using Monte Carlo methods to simulate thousands of game outcomes
- Implemented efficient game state representation and move generation algorithms
- Built frontend with modern JavaScript frameworks for smooth, responsive user experience
- Optimized performance to handle complex calculations without blocking UI
Impact: Created a tool that helps players understand strategic decision-making in Uno through quantitative analysis, making probability theory accessible and engaging.
Big 10 Transition Analysis: Predictive Modeling in College Football
Repository: github.com/bgwarren1/bsa-big10-predictive-analysis
A comprehensive data science project analyzing the 2024 Big Ten conference expansion, predicting how UCLA, USC, Oregon, and Washington would perform after transitioning from the Pac-12.
Research Questions:
- How do environmental factors (weather, travel distance) impact team performance?
- Can we predict win probability based on historical patterns and new conference dynamics?
- What competitive advantages or disadvantages do West Coast teams face in the Big Ten?
Methodology:
- Data Collection: Aggregated multi-season data including game results, weather conditions, travel distances, and team statistics
- Feature Engineering: Created novel metrics combining geographic, climatic, and performance variables
- Predictive Modeling: Built and compared multiple models (logistic regression, random forest, gradient boosting) to forecast game outcomes
- Statistical Validation: Conducted rigorous hypothesis testing and cross-validation to ensure model reliability
Key Findings:
- Quantified the impact of travel fatigue on away game performance for West Coast teams
- Identified weather adaptation as a significant factor in predicted win rates
- Developed conference-adjusted power rankings based on historical matchup data
Technologies Used: Python (pandas, scikit-learn, statsmodels), R (ggplot2, dplyr), Jupyter Notebooks, statistical hypothesis testing
Outcome: Produced actionable insights into competitive dynamics of conference realignment, demonstrating ability to apply data science to sports analytics and strategic forecasting.
Additional Work
I’m continuously working on new projects that explore machine learning, statistical modeling, and data visualization. Check my GitHub for the latest updates and contributions.
Areas of Interest:
- Financial modeling and quantitative trading strategies
- Sports analytics and predictive modeling
- Machine learning applications in real-world domains
- Data visualization and interactive web applications
Want to collaborate or discuss these projects? Get in touch!