Epic Games Tag Mapping

This repository allows to compute an embedding of user tags, based on games on the Epic Games store.

Definition

A "tag" is a word or expression attached to a game, as can be seen on the store page of the game "Fortnite".

Requirements

• Install the latest version of Python 3.X.
• Install the required packages:

pip install -r requirements.txt

Data source

The file data/page_mappings.json:

• contains the mapping between page slugs and sandbox IDs,
• was copied from woctezuma/epic-games-tracker.

Usage

To download tags to data/poll_summary.json and data/poll_dictionary.json, run:

python download_tags.py

To analyze the joint occurences of tags for each game, run:

python map_tags.py

To create a mapping of tags, run epic_games_tag_mapping.ipynb

Results

A 2D embedding is obtained by Uniform Manifold Approximation and Projection (UMAP) with the cosine metric.

With the incidence matrix

This is the map of tags obtained with the incidence matrix as input.

With the adjacency matrix

This is the map of tags obtained with the adjacency matrix as input.

References

• Steam Tag Mapping
• A tracker for ratings and tags (called "awards" there) on the Epic Games store
• A tracker for ratings and achievements