League Of Legends Team Compositions Dataset

Context

League of Legends is one of the most popular videogame nowadays, why? Well, it has everything! Dragons, magic, warriors and, monsters. The game is based on 2 teams composed of 5 players each faced off to destroy the other's base. Each player selects one of the 150 available champions (so far 03/09/2020) who have different abilities and you select it thinking on a specific role. Because of this, one of the most important things in this game is your team's composition. A very interesting question that I want to answer is; Is it possible to predict the result only knowing the two team's composition? (Don't think about the items, winrate or tier) What do you think? 😏

Content

The dataset is composed of 3 files.* Lol_matchs.csv* is the main file containing a table of (34678, 302), the rows represent a match and the 302 columns represent two times the available champions (150*2) plus the match result (a column for each team).The data is already hot-encoded, therefore if a champion's column has a value of 1, means that that champion was used. On the other hand, if you find a 0, it wasn't used. The last 2 columns represent the result. If there is a 1 on the first column means that the first team had the victory, if there is a 0 then means that the other team had the victory. The same rules apply to the last column. Please note that is not possible to have a 1 for both columns and the very first column is just the match ID.
The other 2 files are pickle files. Contains the same information but is separated, Champs are the champions used for each match and Results are the results.
Note: Remember that you can load the pickle data using:

import pickle
with open("Champs", "rb") as f:
Data_champs = pickle.load(f)
(The same applies for the Results file)

Acknowledgements

It took me some time to write the script for getting all this data, so if you like it please give it an upvote.

Inspiration

Let's play LoL but with data 🙌