vchirinosb / call-of-duty Goto Github PK

View Code? Open in Web Editor NEW

Dataset evaluation for Call of Duty. The objective of the case is to use an unsupervised learning technique to cluster players based on their skills. The grouping is by using the clustering method.

Jupyter Notebook 100.00%

call-of-duty's Introduction

Project: Call of Duty

Project description and origin of the dataset

Currently the video game Call of Duty Modern Warfare is one of the most popular online games in the world.
To date there is a problem regarding the types of players that can join in a game (lobby) because there may be players with high skills along with players with low skills.
This inconvenience generates that many new players who may have a low skill level end up not enjoying the game experience because more skilled players eliminate them quickly.
In addition, high skill players are always looking for higher challenges and playing with noobs does not allow them to improve.
The manufacturer Activision also needs to know what types of players it has in general in order to be able to offer different products derived from the game.
Because of this, Activision is looking for a solution to create lobbies where players with similar skills meet.
In this case, a dataset with the game data of more than 1500 players will be used to perform an unsupervised learning process. The data includes number of wins, losses, kill ratio, deaths, assists, among others.
The dataset to be used was found in the Kaggle platform.
Path of the dataset: https://www.kaggle.com/datasets/aishahakami/call-of-duty-players

File: COD.ipynb

Dataset: cod.csv

Python Libraries: pandas, numpy, matplotlib, seaborn, sklearn, mpl_toolkits.

Techniques applied: K-Means.

Recommendations

Run the notebook in Google Colab, without installing python packages on the computer, exploring the available notebook: https://colab.research.google.com/github/vchirinosb/call-of-duty/