This is the final project of CS114 - Machine Learning @ UIT

In this project, we created a collaborative filtering model based on the matrix factorization model, and a content-based filtering recommender based on the softmax model using the data we collected from Steam. This project is mostly based on this course by Google.

You should run the scripts in a Python virtual environment. Create venv/ in the project root using:

python -m venv venv

Activate the virtual environment, then run pip install -r requirements.txt to install the dependencies.

Before doing anything, you should read the README in common/.

First, you will need to scrape the app_ids from Steam:

python -m data.scraper

This script will create a text file called games.txt (specified in common/config.py) containing the scraped app_ids.

Then, collect the games data using:

python -m data.collector collect_games_data

And finally, collect the user reviews data with:

python -m data.collector collect_users_games_data

Sorry for the weird name, it will be changed in the future.

After running the scripts, you should see games.txt, games.csv, and user_game.csv located in data/collected/.

The collected data can be found here.

NOTE: We only collect publicly available data from Steam. Our collected data does not contain any data associated with Steam users, except their steam_id.