With machine learning, the computer is able to cluster similar games together, based on the review patterns that it distills from the review data. For example, maybe some fantasy games exist that are rated high by people who like fantasy games. The computer figures out wat 'similar' means with only the users, games and ratings at it's disposal. With this site you can also explore the results!
For the enthusiasts who want to know what's under the engine: At Kaggle I've uploaded the data, including some exploratory data analysis and an in-depth analysis. And the code of the app on GitHub.
Try it out yourself and have fun!
|Minimum number of ratings for a game|
|Number of similar games in the result|