(64) Analyzing player reviews with natural language processing to identify ecogames for education and research

Ecogames – games that address environmental topics – are an innovative and engaging method for fostering environmental education, which is urgently needed to respond effectively to climate change. However, educators and researchers are challenged by the task of selecting and investigating appropriate games from the vast and ever-growing assortment available. The rapid pace and large volume of the video game market complicate the selection process for classroom activities or research, such as media comparison studies. This study introduces a novel methodology that employs natural language processing techniques to identify and preselect ecogames that feature climate change-related content to varying degrees. By analyzing over 186,000 player-generated reviews on the Steam gaming platform through dictionary-based analysis and climate change detection classification, the methodology effectively distinguishes ecogames from action/adventure games and demonstrates varying levels of climate change-related content across different game categories. It also highlights that not all ecogames feature discussion on environmental matters equally, underscoring the need for careful game selection. Additionally, this approach has uncovered new ecogame candidates among a random selection of action/adventure games. Relying solely on the lexical features of the texts investigated, this study’s approach is quick and scalable in its implementation and represents a data-driven way of preselecting ecogames based on publicly available text data. Combining the text analyses with other game metrics such as playtime, game popularity and age rating, the study recognizes Beecarbonize, Alba: A Wildlife Adventure and Gibbon: Beyond the Trees as notable candidates for in-depth research and potential educational use in the classroom.

Recommended citation: Edlinger, M., Huber, S. E., & Ninaus, M. (2025). Analyzing player reviews with natural language processing to identify ecogames for education and research. Computers in Human Behavior Reports, 20, 100850.
Get Paper