Representação da informação de jogos digitais na Plataforma Steam

Abstract

The Steam platform stands out as one of the main and most dynamic platforms for digital games today, driven by an active and highly engaged community. On Steam, the community plays a key role even in influencing indexing terms through the tagging system. Despite this engagement, the platform's search and recommendation system is not always accurate, due to its light hierarchical nature and poorly defined relationships within the categorization system. This work identified opportunities for improvement in this model and proposed strategies to improve the indexing and information retrieval processes. The proposal consists of a hybrid representation model, aiming to optimize the way data is organized and retrieved. The research adopted an applied, qualitative and exploratory approach, based on content analysis. The way information is represented on Steam was investigated, culminating in the creation of a thesaurus. This thesaurus served as the basis for proposing an assisted folksonomy model, adapted to be applied on the platform. The study included the collection and categorization of markers present in Steam’s documentation, as well as hierarchization through the thesaurus. The results indicate the possibility of linking this folksonomy model to the existing structure on Steam, thus forming an assisted folksonomy. The application of these models revealed their complementarity and suitability to the needs of the Steam platform users, suggesting the feasibility of implementing a more robust and efficient system.