All projects
•
2 min read

SteamMatcher

Recommender system

python pandas scikit-learn docker fastapi
Table of Contents

Introduction

This is a college project I did alongside my wonderful classmates.

The model recommends games from Steam, an online video game store. It applies a collaborative filtering technique based on building graphs with relationships between games using community data and then suggest titles similar to those in a player’s profile.

The source code is not hosted public. Contact me if you want access.

Web Interface

There is a web interface at the root (/) that allows the user to interact with the model.

This is the first screen the user sees. It lets the user enter a Steam user ID to search.

SteamMatcher - main menu

Error menu

This screen appears if an error occurs while fetching the user’s information, for example if the profile is private or the ID is invalid.

SteamMatcher - error menu

Recommendations menu

This screen shows the 20 recommendations returned for the user. It includes the game cover, its price, the most relevant user tags, and a short description. By clicking the images, you can go directly to the game’s store page on Steam.

SteamMatcher - recommendations menu

REST API

There is an API endpoint that returns a JSON file containing the top 20 recommendations for a given user.

Endpoint

GET /recommendation/json

Co-Authors

  • Daniel Pantoja Joaristi
  • Jorge Hernández Palop
  • Marco Casteleiro Trigueros
  • Álvaro Arbona Rodríguez
  • Óscar Marín Esteban