This research paper deals with the video game recommendation system on the Steam platform. The goal is to recommend games to users that are similar in certain parameters but not focusing only on popular games. In addition, a detailed Exploratory Data Analysis (EDA) was conducted to find out which factors influence the popularity of the game and which games the players prefer the most. It is found that some of these factors are game genres, prices, ratings, and publisher reputation. Very important insights into the popularity of games and the connection between different features were gained through the testing of hypotheses. To provide better recommendations to users, advanced metrics and custom models of K-nearest neighbors (KNN) have been developed. The biggest advantage of this system is that it provides a balance between recommending games that are relevant based on features and at the same time provides a different variety of games. This approach can be very useful in the further creation of advanced recommendation systems and in addition offers very significant insights into the video game industry itself, which can be very effective in improving user engagement and satisfaction on a platform such as Steam.

Steam video game recommendation Exploratory Data Analysis (EDA) content-based recommender system K-Nearest Neighbors (KNN) game popularity player preferences user engagement analytical framework