Portfolio
Esports Analytics Projects
Please click on the images to see more!
Draft Master Sheet System
I created a comprehensive draft tracking and analysis system in Google Sheets that automatically integrates with my database backend. This tool provides coaches and analysts with complete historical draft information in an easily accessible format. The system features vertical draft views for intuitive pattern recognition, automatically computed first pick/ban statistics, win rate breakdowns by side and pick order, and champion usage trends. This toolkit has become essential for our pre-match preparation on Luminosity Gaming.
Champion Priority Analysis
I furthered the work of Beora, combining player-specific performance data with meta trends to generate champion priority recommendations. The system employs weighted temporal analysis that values recent games higher than older ones and applies role-specific meta-influence factors to try and predict the champions a player is likely to play in future matches. I would love to continue developing more tools like this using the available data.
Solo Queue Tracker
This tracker displays data in an easy-to-view format, allowing coaches and analysts to check what players practice in solo queue and easily scout opponents. On the backend, it's tied into my main game database, so it was trivial to begin scouting our opponents for the Americas tournament even though our opponents were playing on boot camp accounts and were originally from a different region.
Automated Visualizations - Lvl 1 Scouting, Jungle Maps, Heatmaps
Here are some examples of visualizations that I worked on while with AOE. A lot of things broke during the transition from Bayes to GRID, these map data based tools being one of them, but it wouldn't be too hard to revisit.
Software Development Projects
ATG - All The Games
ATG (All The Games) is a high-performance game data management system I've open sourced to unifify data from scrims, matches, and solo queue games. Using Python, SQLAlchemy, and PostgreSQL, I designed flexible database models that store complete game data using JSONB columns for maximum adaptability. This allows us to track over 310k games and over 41 million game events! (April 2025). The database models allow for easy analysis of various custom stats!
lol_bayes_auth
I developed a Python package for the Bayes Esports API, simplifying the obtaining and managing authentication tokens. The package applies caching and automatic refresh mechanisms, reducing unnecessary authentication requests. I implemented comprehensive unit tests using pytest, requests-mock, and freeze gun to ensure code reliability and ease of maintenance.
Automated Analysis Pipeline
I built a fully automated analysis pipeline that continuously updates all team tools and databases with minimal human intervention. The system runs hourly data collection jobs, scheduled daily priority score calculations. This automation ensures coaches always have the latest information available. The pipeline includes error handling, logging, and notification systems to ensure reliable operation, with 95% uptime since deployment.