This is a project for my Programming 1 class.
In this project I have analysed a few thousand best selling games from the online platform Steam.
For each game I have collected the following data:
- Name and description of the game
- Release date
- Metascore and age rating
- Genres of the game
- Number of all reviews and the amount of positive reviews
- Number of recent reviews and the amount of positive recent reviews
- Developer and publisher
- Price of the game
- Bundles in which the game is available
- Available additional content for the games
- Game attributes
- Minimum and reccomended system requirements
A few of the questions I have tried to answer are:
- Is there a correlation between the price of the game and its publisher?
- Which developers/publishers have the best rated games?
- Are people more likely to review expensive games?
- How much more computationaly demanding are games becoming with time?
- Is it possible to predict game's tags using it's description?
From the data I have concluded that games are becoming more and more computationaly demanding with time, in accordance with Moore's Law. Based on the user reviews, the developer most consistently releasing good games is Fireproof games
, with the series The Room
.
It turned out that predicting tags from the game's description is very difficult, most likely because there is apparently almost no relation between the two.