Player Performance Analysis

This study delved into professional League of Legends (LoL) match data from 2022, sourced from Oracle’s Elixir. The primary objective was to investigate whether a player's versatility in champion selection correlates with their success rate in matches.​

-Data Overview: The dataset comprised 150,180 entries with 161 attributes each, detailing various aspects of gameplay, individual performances, and team dynamics.​

-Data Cleaning: Focused on extracting relevant columns such as player roles, names, champions used, and match outcomes. The dataset was filtered to include only player-specific information, ensuring accuracy by removing entries with missing values.​

-Feature Engineering: Introduced metrics like the number of unique champions played (num_characters) and a weighted diversity measure (weighted_num_characters) using Shannon diversity. Players with fewer than 20 matches were excluded to maintain statistical significance.​

-Analysis: Calculated each player's win rate and examined its relationship with their champion diversity. The data was sorted based on the weighted number of characters to identify patterns.