Skip to content
FA550 Data Visualization · Stevens Institute of Technology · Fall 2025

The Steam game ecosystem

I started by comparing game genres, prices, VR and language support on Steam. I became more interested in the growth of independent creators, and who gets to play and understand their games.

More people can make games outside large studios now. I wanted to see whether those games are also becoming accessible to people who speak different languages.

Author
Zhenhan Chen
Data
Kaggle Steam store and player datasets
Tools
Python · pandas · Tableau
Presentation
Original Tableau exports with commentary from my presentation notes

Original Tableau charts from my presentation. Select a chart to view the full-size image. These exports are static.

01

Independent creators

Indie describes how a game is made, rather than a genre.

A game can be both indie and action, adventure or simulation. I separated the indie tag from genre when comparing the catalogue. These charts count games, not players or sales.

Game counts by genre after excluding Indie and software categories
Genre counts with the Indie tag and software categories excluded.
Original Tableau chart showing the growth of indie releases from 2005 to 2012
The early rise of indie releases, shown in the original presentation’s 2005–2012 view.

The rise of indie games reminded me of what happened to publishing, photography and video. Work that once needed a large organization could increasingly be made by one person or a small team. I see this as a change in who gets to express themselves through games, rather than simply another category becoming popular.

The release counts describe the supply of games, rather than their audience or sales.

Average review scores by genre and release year
Review scores by release year and genre. The presentation discusses the recovery in Early Access ratings and the lower scores of massively multiplayer games.

I also compared mean review scores. Casual games scored 12.86 points above massively multiplayer games, and indie games 11.85 points above them, with reported p-values below 0.001. These comparisons describe the groups in this dataset; they do not establish that a smaller team makes a better game.

02

Price and player ratings

A higher price does not consistently correspond to a higher review score.

Average game price by release year
Average game price grouped by release year, as shown in Tableau.
Average review scores by price category
Review scores by price category. The $15–30 category has the highest average in this comparison.
Review scores over time for low, mid and high price categories
Review scores over time for three price categories, including the changing position of lower-priced games.

In the talk, I suggested that players might expect more from an expensive game, which could affect their ratings. That is a possible explanation, not something the charts test. The price-by-release-year chart also is not a history of what the same games cost at launch, and it does not by itself support an inflation-adjusted price comparison.

03

VR games

More VR releases, higher average prices and lower average ratings in this dataset.

Number of VR games by release year
VR releases increase after 2013 in the original chart.
Average prices for VR and non-VR games
The original comparison of average prices for VR and non-VR games.
Average review scores for VR and non-VR games
VR games receive lower average review scores in this comparison.

I was skeptical that the higher price of VR games was reflected in the experience players reported. It also made me question how much of the discussion about the future of games should focus on new hardware. The growth of independent development seemed more consequential to me.

04

Language and access

Who can understand a game is part of who can experience it.

Number of games supporting each language
English is the most widely supported language, followed by Simplified Chinese in this dataset.
Number of supported languages per game
Most games support relatively few languages.
Annual counts of English-only, multilingual and single non-English releases
Annual release counts by language category. The categories distinguish English-only, multilingual and single non-English games.
Shares of English-only, multilingual and single non-English games by release year
Language categories as a share of releases. This is the proportion view from the presentation, rather than the release-count view.

I then separated indie and non-indie games to look at the rise in English-only releases around 2014. The increase was especially visible in indie releases. The non-indie comparison followed a different pattern. The count and proportion charts help separate these changes.

Counts and shares of language support categories among indie games
Indie games: release counts above, language category shares below.
Counts and shares of language support categories among non-indie games
Non-indie games: the corresponding count and share views.

Translation takes work, and in the talk I speculated about developers balancing that work against reaching more players. The dataset does not include translation budgets or developer intentions. My conclusion was about access: more language support gives more people a chance to understand what a developer wants to say.

05

Why this matters to me

When I came to the United States, video games were one of the main ways I made friends with Americans. Games gave us something to do together even when we came from different backgrounds.

I think of games as an art form. They carry what their creators want to say, and players take part through interaction. Independent development gives more people a way to make that work. Supporting more languages gives more people a way to experience it.

I am happy to see more games supporting multiple languages. I want more people to be able to understand them and play together.