OptiDraft

A champion drafting and recommender system for League of Legends

Quick Summary

  • Spearheaded a 5-member team in architecting a character recommender system for League of Legends
  • Built a robust ETL pipeline processing 36K API-sourced match records with fault tolerance and memory optimization
  • Created synergy metrics for team dynamics, achieving 20% win-rate improvement through fine-tuned models (logistic regression, random forest, MLP)
  • Launched public web interface providing real-time draft suggestions

Tools: Python (Scikit-learn, Pandas, Matplotlib), PostgreSQL, Jupyter Notebook, GitHub

Web Interface

For the best user experience, please open the link on a desktop: https://optidraft.github.io/optidraft/
For more detailed project information, please visit the GitHub repo: https://github.com/optidraft/lol-draft-pick