Steam Hybrid Graph Recommender: Sampled Interaction Dataset and Enriched Metadata

This dataset contains preprocessed and enriched data for the graph-based hybrid recommendation system developed for the Algorithms and Data Structures II course at CEFET-MG. Files included:- sampled_recommendations.csv: Bipartite graph interaction edges filtered via adaptive K-Core based on median playtime hours [10h, 50h].- games_metadata.json: Enriched metadata containing top 20 community tags, official Steam genres, and normalized developer studios retrieved via SteamSpy API.- games.csv: Master mapping of Steam App IDs to game titles and review statistics. Source raw data origin: Kaggle (Game Recommendations on Steam by Anton Kozyriev).Code repository: https://github.com/rafa16w34/TrabalhoFinal-AEDS2